Generate sha256 hashes when packaging

(cherry picked from commit 0b6db905ff)

.github/ISSUE_TEMPLATE.md

@@ -1,7 +1,7 @@
 ## Directions
 
 GitHub Issues are reserved for actionable bug reports and feature requests.
-General questions should be sent to the [InfluxDB mailing list](https://groups.google.com/forum/#!forum/influxdb).
+General questions should be asked at the [InfluxData Community](https://community.influxdata.com) site.
 
 Before opening an issue, search for similar bug reports or feature requests on GitHub Issues.
 If no similar issue can be found, fill out either the "Bug Report" or the "Feature Request" section below.

CHANGELOG.md

@@ -1,15 +1,225 @@
-## v1.2 [unreleased]
-
-### Release Notes
-
-### Features
+## v1.3.1 [unreleased]
 
 ### Bugfixes
 
-## v1.1 [unreleased]
-- [#1949](https://github.com/influxdata/telegraf/issues/1949): Fix windows `net` plugin.
+- [#2749](https://github.com/influxdata/telegraf/pull/2749): Fixed sqlserver input to work with case sensitive server collation.
+- [#2782](https://github.com/influxdata/telegraf/pull/2782): Reuse transports in input plugins
+- [#2815](https://github.com/influxdata/telegraf/issues/2815): Inputs processes fails with "no such process".
+- [#2851](https://github.com/influxdata/telegraf/pull/2851): Fix InfluxDB output database quoting.
+- [#2856](https://github.com/influxdata/telegraf/issues/2856): Fix net input on older Linux kernels.
+- [#2848](https://github.com/influxdata/telegraf/pull/2848): Fix panic in mongo input.
+- [#2869](https://github.com/influxdata/telegraf/pull/2869): Fix length calculation of split metric buffer.
+
+## v1.3 [2017-05-15]
+
+### Release Notes
+
+- Users of the windows `ping` plugin will need to drop or migrate their
+measurements in order to continue using the plugin. The reason for this is that
+the windows plugin was outputting a different type than the linux plugin. This
+made it impossible to use the `ping` plugin for both windows and linux
+machines.
+
+- Ceph: the `ceph_pgmap_state` metric content has been modified to use a unique field `count`, with each state expressed as a `state` tag.
+
+Telegraf < 1.3:
+
+```
+# field_name             value
+active+clean             123
+active+clean+scrubbing   3
+```
+
+Telegraf >= 1.3:
+
+```
+# field_name    value       tag
+count           123         state=active+clean
+count           3           state=active+clean+scrubbing
+```
+
+- The [Riemann output plugin](./plugins/outputs/riemann) has been rewritten
+and the previous riemann plugin is _incompatible_ with the new one. The reasons
+for this are outlined in issue [#1878](https://github.com/influxdata/telegraf/issues/1878).
+The previous riemann output will still be available using
+`outputs.riemann_legacy` if needed, but that will eventually be deprecated.
+It is highly recommended that all users migrate to the new riemann output plugin.
+
+- Generic [socket_listener](./plugins/inputs/socket_listener) and
+[socket_writer](./plugins/outputs/socket_writer) plugins have been implemented
+for receiving and sending UDP, TCP, unix, & unix-datagram data. These plugins
+will replace udp_listener and tcp_listener, which are still available but will
+be deprecated eventually.
+
+### Features
+
+- [#2721](https://github.com/influxdata/telegraf/pull/2721): Added SASL options for kafka output plugin.
+- [#2723](https://github.com/influxdata/telegraf/pull/2723): Added SSL configuration for input haproxy.
+- [#2494](https://github.com/influxdata/telegraf/pull/2494): Add interrupts input plugin.
+- [#2094](https://github.com/influxdata/telegraf/pull/2094): Add generic socket listener & writer.
+- [#2204](https://github.com/influxdata/telegraf/pull/2204): Extend http_response to support searching for a substring in response. Return 1 if found, else 0.
+- [#2137](https://github.com/influxdata/telegraf/pull/2137): Added userstats to mysql input plugin.
+- [#2179](https://github.com/influxdata/telegraf/pull/2179): Added more InnoDB metric to MySQL plugin.
+- [#2229](https://github.com/influxdata/telegraf/pull/2229): `ceph_pgmap_state` metric now uses a single field `count`, with PG state published as `state` tag.
+- [#2251](https://github.com/influxdata/telegraf/pull/2251): InfluxDB output: use own client for improved through-put and less allocations.
+- [#2330](https://github.com/influxdata/telegraf/pull/2330): Keep -config-directory when running as Windows service.
+- [#1900](https://github.com/influxdata/telegraf/pull/1900): Riemann plugin rewrite.
+- [#1453](https://github.com/influxdata/telegraf/pull/1453): diskio: add support for name templates and udev tags.
+- [#2277](https://github.com/influxdata/telegraf/pull/2277): add integer metrics for Consul check health state.
+- [#2201](https://github.com/influxdata/telegraf/pull/2201): Add lock option to the IPtables input plugin.
+- [#2244](https://github.com/influxdata/telegraf/pull/2244): Support ipmi_sensor plugin querying local ipmi sensors.
+- [#2339](https://github.com/influxdata/telegraf/pull/2339): Increment gather_errors for all errors emitted by inputs.
+- [#2071](https://github.com/influxdata/telegraf/issues/2071): Use official docker SDK.
+- [#1678](https://github.com/influxdata/telegraf/pull/1678): Add AMQP consumer input plugin
+- [#2512](https://github.com/influxdata/telegraf/pull/2512): Added pprof tool.
+- [#2501](https://github.com/influxdata/telegraf/pull/2501): Support DEAD(X) state in system input plugin.
+- [#2522](https://github.com/influxdata/telegraf/pull/2522): Add support for mongodb client certificates.
+- [#1948](https://github.com/influxdata/telegraf/pull/1948): Support adding SNMP table indexes as tags.
+- [#2332](https://github.com/influxdata/telegraf/pull/2332): Add Elasticsearch 5.x output
+- [#2587](https://github.com/influxdata/telegraf/pull/2587): Add json timestamp units configurability
+- [#2597](https://github.com/influxdata/telegraf/issues/2597): Add support for Linux sysctl-fs metrics.
+- [#2425](https://github.com/influxdata/telegraf/pull/2425): Support to include/exclude docker container labels as tags
+- [#1667](https://github.com/influxdata/telegraf/pull/1667): dmcache input plugin
+- [#2637](https://github.com/influxdata/telegraf/issues/2637): Add support for precision in http_listener
+- [#2636](https://github.com/influxdata/telegraf/pull/2636): Add `message_len_max` option to `kafka_consumer` input
+- [#1100](https://github.com/influxdata/telegraf/issues/1100): Add collectd parser
+- [#1820](https://github.com/influxdata/telegraf/issues/1820): easier plugin testing without outputs
+- [#2493](https://github.com/influxdata/telegraf/pull/2493): Check signature in the GitHub webhook plugin
+- [#2038](https://github.com/influxdata/telegraf/issues/2038): Add papertrail support to webhooks
+- [#2253](https://github.com/influxdata/telegraf/pull/2253): Change jolokia plugin to use bulk requests.
+- [#2575](https://github.com/influxdata/telegraf/issues/2575) Add diskio input for Darwin
+- [#2705](https://github.com/influxdata/telegraf/pull/2705): Kinesis output: add use_random_partitionkey option
+- [#2635](https://github.com/influxdata/telegraf/issues/2635): add tcp keep-alive to socket_listener & socket_writer
+- [#2031](https://github.com/influxdata/telegraf/pull/2031): Add Kapacitor input plugin
+- [#2732](https://github.com/influxdata/telegraf/pull/2732): Use go 1.8.1
+- [#2712](https://github.com/influxdata/telegraf/issues/2712): Documentation for rabbitmq input plugin
+- [#2141](https://github.com/influxdata/telegraf/pull/2141): Logparser handles newly-created files.
+
+### Bugfixes
+
+- [#2633](https://github.com/influxdata/telegraf/pull/2633): ipmi_sensor: allow @ symbol in password
+- [#2077](https://github.com/influxdata/telegraf/issues/2077): SQL Server Input - Arithmetic overflow error converting numeric to data type int.
+- [#2262](https://github.com/influxdata/telegraf/issues/2262): Flush jitter can inhibit metric collection.
+- [#2318](https://github.com/influxdata/telegraf/issues/2318): haproxy input - Add missing fields.
+- [#2287](https://github.com/influxdata/telegraf/issues/2287): Kubernetes input: Handle null startTime for stopped pods.
+- [#2356](https://github.com/influxdata/telegraf/issues/2356): cpu input panic when /proc/stat is empty.
+- [#2341](https://github.com/influxdata/telegraf/issues/2341): telegraf swallowing panics in --test mode.
+- [#2358](https://github.com/influxdata/telegraf/pull/2358): Create pidfile with 644 permissions & defer file deletion.
+- [#2360](https://github.com/influxdata/telegraf/pull/2360): Fixed install/remove of telegraf on non-systemd Debian/Ubuntu systems
+- [#2282](https://github.com/influxdata/telegraf/issues/2282): Reloading telegraf freezes prometheus output.
+- [#2390](https://github.com/influxdata/telegraf/issues/2390): Empty tag value causes error on InfluxDB output.
+- [#2380](https://github.com/influxdata/telegraf/issues/2380): buffer_size field value is negative number from "internal" plugin.
+- [#2414](https://github.com/influxdata/telegraf/issues/2414): Missing error handling in the MySQL plugin leads to segmentation violation.
+- [#2462](https://github.com/influxdata/telegraf/pull/2462): Fix type conflict in windows ping plugin.
+- [#2178](https://github.com/influxdata/telegraf/issues/2178): logparser: regexp with lookahead.
+- [#2466](https://github.com/influxdata/telegraf/issues/2466): Telegraf can crash in LoadDirectory on 0600 files.
+- [#2215](https://github.com/influxdata/telegraf/issues/2215): Iptables input: document better that rules without a comment are ignored.
+- [#2483](https://github.com/influxdata/telegraf/pull/2483): Fix win_perf_counters capping values at 100.
+- [#2498](https://github.com/influxdata/telegraf/pull/2498): Exporting Ipmi.Path to be set by config.
+- [#2500](https://github.com/influxdata/telegraf/pull/2500): Remove warning if parse empty content
+- [#2520](https://github.com/influxdata/telegraf/pull/2520): Update default value for Cloudwatch rate limit
+- [#2513](https://github.com/influxdata/telegraf/issues/2513): create /etc/telegraf/telegraf.d directory in tarball.
+- [#2541](https://github.com/influxdata/telegraf/issues/2541): Return error on unsupported serializer data format.
+- [#1827](https://github.com/influxdata/telegraf/issues/1827): Fix Windows Performance Counters multi instance identifier
+- [#2576](https://github.com/influxdata/telegraf/pull/2576): Add write timeout to Riemann output
+- [#2596](https://github.com/influxdata/telegraf/pull/2596): fix timestamp parsing on prometheus plugin
+- [#2610](https://github.com/influxdata/telegraf/pull/2610): Fix deadlock when output cannot write
+- [#2410](https://github.com/influxdata/telegraf/issues/2410): Fix connection leak in postgresql.
+- [#2628](https://github.com/influxdata/telegraf/issues/2628): Set default measurement name for snmp input.
+- [#2649](https://github.com/influxdata/telegraf/pull/2649): Improve performance of diskio with many disks
+- [#2671](https://github.com/influxdata/telegraf/issues/2671): The internal input plugin uses the wrong units for `heap_objects`
+- [#2684](https://github.com/influxdata/telegraf/pull/2684): Fix ipmi_sensor config is shared between all plugin instances
+- [#2450](https://github.com/influxdata/telegraf/issues/2450): Network statistics not collected when system has alias interfaces
+- [#1911](https://github.com/influxdata/telegraf/issues/1911): Sysstat plugin needs LANG=C or similar locale
+- [#2528](https://github.com/influxdata/telegraf/issues/2528): File output closes standard streams on reload.
+- [#2603](https://github.com/influxdata/telegraf/issues/2603): AMQP output disconnect blocks all outputs
+- [#2706](https://github.com/influxdata/telegraf/issues/2706): Improve documentation for redis input plugin
+
+## v1.2.1 [2017-02-01]
+
+### Bugfixes
+
+- [#2317](https://github.com/influxdata/telegraf/issues/2317): Fix segfault on nil metrics with influxdb output.
+- [#2324](https://github.com/influxdata/telegraf/issues/2324): Fix negative number handling.
+
+### Features
+
+- [#2348](https://github.com/influxdata/telegraf/pull/2348): Go version 1.7.4 -> 1.7.5
+
+## v1.2 [2017-01-00]
+
+### Release Notes
+
+- The StatsD plugin will now default all "delete_" config options to "true". This
+will change te default behavior for users who were not specifying these parameters
+in their config file.
+
+- The StatsD plugin will also no longer save it's state on a service reload.
+Essentially we have reverted PR [#887](https://github.com/influxdata/telegraf/pull/887).
+The reason for this is that saving the state in a global variable is not
+thread-safe (see [#1975](https://github.com/influxdata/telegraf/issues/1975) & [#2102](https://github.com/influxdata/telegraf/issues/2102)),
+and this creates issues if users want to define multiple instances
+of the statsd plugin. Saving state on reload may be considered in the future,
+but this would need to be implemented at a higher level and applied to all
+plugins, not just statsd.
+
+### Features
+
+- [#2123](https://github.com/influxdata/telegraf/pull/2123): Fix improper calculation of CPU percentages
+- [#1564](https://github.com/influxdata/telegraf/issues/1564): Use RFC3339 timestamps in log output.
+- [#1997](https://github.com/influxdata/telegraf/issues/1997): Non-default HTTP timeouts for RabbitMQ plugin.
+- [#2074](https://github.com/influxdata/telegraf/pull/2074): "discard" output plugin added, primarily for testing purposes.
+- [#1965](https://github.com/influxdata/telegraf/pull/1965): The JSON parser can now parse an array of objects using the same configuration.
+- [#1807](https://github.com/influxdata/telegraf/pull/1807): Option to use device name rather than path for reporting disk stats.
+- [#1348](https://github.com/influxdata/telegraf/issues/1348): Telegraf "internal" plugin for collecting stats on itself.
+- [#2127](https://github.com/influxdata/telegraf/pull/2127): Update Go version to 1.7.4.
+- [#2126](https://github.com/influxdata/telegraf/pull/2126): Support a metric.Split function.
+- [#2026](https://github.com/influxdata/telegraf/pull/2065): elasticsearch "shield" (basic auth) support doc.
+- [#1885](https://github.com/influxdata/telegraf/pull/1885): Fix over-querying of cloudwatch metrics
+- [#1913](https://github.com/influxdata/telegraf/pull/1913): OpenTSDB basic auth support.
+- [#1908](https://github.com/influxdata/telegraf/pull/1908): RabbitMQ Connection metrics.
+- [#1937](https://github.com/influxdata/telegraf/pull/1937): HAProxy session limit metric.
+- [#2068](https://github.com/influxdata/telegraf/issues/2068): Accept strings for StatsD sets.
+- [#1893](https://github.com/influxdata/telegraf/issues/1893): Change StatsD default "reset" behavior.
+- [#2079](https://github.com/influxdata/telegraf/pull/2079): Enable setting ClientID in MQTT output.
+- [#2001](https://github.com/influxdata/telegraf/pull/2001): MongoDB input plugin: Improve state data.
+- [#2078](https://github.com/influxdata/telegraf/pull/2078): Ping input: add standard deviation field.
+- [#2121](https://github.com/influxdata/telegraf/pull/2121): Add GC pause metric to InfluxDB input plugin.
+- [#2006](https://github.com/influxdata/telegraf/pull/2006): Added response_timeout property to prometheus input plugin.
+- [#1763](https://github.com/influxdata/telegraf/issues/1763): Pulling github.com/lxn/win's pdh wrapper into telegraf.
+- [#1898](https://github.com/influxdata/telegraf/issues/1898): Support negative statsd counters.
+- [#1921](https://github.com/influxdata/telegraf/issues/1921): Elasticsearch cluster stats support.
+- [#1942](https://github.com/influxdata/telegraf/pull/1942): Change Amazon Kinesis output plugin to use the built-in serializer plugins.
+- [#1980](https://github.com/influxdata/telegraf/issues/1980): Hide username/password from elasticsearch error log messages.
+- [#2097](https://github.com/influxdata/telegraf/issues/2097): Configurable HTTP timeouts in Jolokia plugin
+- [#2255](https://github.com/influxdata/telegraf/pull/2255): Allow changing jolokia attribute delimiter
+
+### Bugfixes
+
+- [#2049](https://github.com/influxdata/telegraf/pull/2049): Fix the Value data format not trimming null characters from input.
+- [#1949](https://github.com/influxdata/telegraf/issues/1949): Fix windows `net` plugin.
+- [#1775](https://github.com/influxdata/telegraf/issues/1775): Cache & expire metrics for delivery to prometheus
+- [#1775](https://github.com/influxdata/telegraf/issues/1775): Cache & expire metrics for delivery to prometheus.
+- [#2146](https://github.com/influxdata/telegraf/issues/2146): Fix potential panic in aggregator plugin metric maker.
+- [#1843](https://github.com/influxdata/telegraf/pull/1843) & [#1668](https://github.com/influxdata/telegraf/issues/1668): Add optional ability to define PID as a tag.
+- [#1730](https://github.com/influxdata/telegraf/issues/1730) & [#2261](https://github.com/influxdata/telegraf/pull/2261): Fix win_perf_counters not gathering non-English counters.
+- [#2061](https://github.com/influxdata/telegraf/issues/2061): Fix panic when file stat info cannot be collected due to permissions or other issue(s).
+- [#2045](https://github.com/influxdata/telegraf/issues/2045): Graylog output should set short_message field.
+- [#1904](https://github.com/influxdata/telegraf/issues/1904): Hddtemp always put the value in the field temperature.
+- [#1693](https://github.com/influxdata/telegraf/issues/1693): Properly collect nested jolokia struct data.
+- [#1917](https://github.com/influxdata/telegraf/pull/1917): fix puppetagent inputs plugin to support string for config variable.
+- [#1987](https://github.com/influxdata/telegraf/issues/1987): fix docker input plugin tags when registry has port.
+- [#2089](https://github.com/influxdata/telegraf/issues/2089): Fix tail input when reading from a pipe.
+- [#1449](https://github.com/influxdata/telegraf/issues/1449): MongoDB plugin always shows 0 replication lag.
+- [#1825](https://github.com/influxdata/telegraf/issues/1825): Consul plugin: add check_id as a tag in metrics to avoid overwrites.
+- [#1973](https://github.com/influxdata/telegraf/issues/1973): Partial fix: logparser CLF pattern with IPv6 addresses.
+- [#1975](https://github.com/influxdata/telegraf/issues/1975) & [#2102](https://github.com/influxdata/telegraf/issues/2102): Fix thread-safety when using multiple instances of the statsd input plugin.
+- [#2027](https://github.com/influxdata/telegraf/issues/2027): docker input: interface conversion panic fix.
+- [#1814](https://github.com/influxdata/telegraf/issues/1814): snmp: ensure proper context is present on error messages.
+- [#2299](https://github.com/influxdata/telegraf/issues/2299): opentsdb: add tcp:// prefix if no scheme provided.
+- [#2297](https://github.com/influxdata/telegraf/issues/2297): influx parser: parse line-protocol without newlines.
+- [#2245](https://github.com/influxdata/telegraf/issues/2245): influxdb output: fix field type conflict blocking output buffer.
 
-## v1.1.1 [unreleased]
 ## v1.1.2 [2016-12-12]
 
 ### Bugfixes
@@ -98,6 +308,7 @@ continue sending logs to /var/log/telegraf/telegraf.log.
 - [#1771](https://github.com/influxdata/telegraf/issues/1771): Delete nil fields in the metric maker.
 - [#870](https://github.com/influxdata/telegraf/issues/870): Fix MySQL special characters in DSN parsing.
 - [#1742](https://github.com/influxdata/telegraf/issues/1742): Ping input odd timeout behavior.
+- [#1950](https://github.com/influxdata/telegraf/pull/1950): Switch to github.com/kballard/go-shellquote.
 
 ## v1.0.1 [2016-09-26]
 
@@ -158,8 +369,11 @@ which can be installed via
 evaluated at every flush interval, rather than once at startup. This makes it
 consistent with the behavior of `collection_jitter`.
 
+- postgresql plugins now handle oid and name typed columns seamlessly, previously they were ignored/skipped.
+
 ### Features
 
+- [#1617](https://github.com/influxdata/telegraf/pull/1617): postgresql_extensible now handles name and oid types correctly.
 - [#1413](https://github.com/influxdata/telegraf/issues/1413): Separate container_version from container_image tag.
 - [#1525](https://github.com/influxdata/telegraf/pull/1525): Support setting per-device and total metrics for Docker network and blockio.
 - [#1466](https://github.com/influxdata/telegraf/pull/1466): MongoDB input plugin: adding per DB stats from db.stats()

CONTRIBUTING.md

@@ -124,7 +124,7 @@ You should also add the following to your SampleConfig() return:
 
 ```toml
   ## Data format to consume.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
   data_format = "influx"
@@ -254,7 +254,7 @@ You should also add the following to your SampleConfig() return:
 
 ```toml
   ## Data format to output.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_OUTPUT.md
   data_format = "influx"

Godeps

@@ -1,65 +1,65 @@
-github.com/Shopify/sarama 8aadb476e66ca998f2f6bb3c993e9a2daa3666b9
-github.com/Sirupsen/logrus 219c8cb75c258c552e999735be6df753ffc7afdc
-github.com/aerospike/aerospike-client-go 7f3a312c3b2a60ac083ec6da296091c52c795c63
-github.com/amir/raidman 53c1b967405155bfc8758557863bf2e14f814687
-github.com/aws/aws-sdk-go 13a12060f716145019378a10e2806c174356b857
-github.com/beorn7/perks 3ac7bf7a47d159a033b107610db8a1b6575507a4
-github.com/cenkalti/backoff 4dc77674aceaabba2c7e3da25d4c823edfb73f99
-github.com/couchbase/go-couchbase cb664315a324d87d19c879d9cc67fda6be8c2ac1
-github.com/couchbase/gomemcached a5ea6356f648fec6ab89add00edd09151455b4b2
+collectd.org 2ce144541b8903101fb8f1483cc0497a68798122
+github.com/Shopify/sarama 574d3147eee384229bf96a5d12c207fe7b5234f3
+github.com/Sirupsen/logrus 61e43dc76f7ee59a82bdf3d71033dc12bea4c77d
+github.com/aerospike/aerospike-client-go 95e1ad7791bdbca44707fedbb29be42024900d9c
+github.com/amir/raidman c74861fe6a7bb8ede0a010ce4485bdbb4fc4c985
+github.com/aws/aws-sdk-go 7524cb911daddd6e5c9195def8e59ae892bef8d9
+github.com/beorn7/perks 4c0e84591b9aa9e6dcfdf3e020114cd81f89d5f9
+github.com/cenkalti/backoff b02f2bbce11d7ea6b97f282ef1771b0fe2f65ef3
+github.com/couchbase/go-couchbase bfe555a140d53dc1adf390f1a1d4b0fd4ceadb28
+github.com/couchbase/gomemcached 4a25d2f4e1dea9ea7dd76dfd943407abf9b07d29
 github.com/couchbase/goutils 5823a0cbaaa9008406021dc5daf80125ea30bba6
-github.com/dancannon/gorethink e7cac92ea2bc52638791a021f212145acfedb1fc
-github.com/davecgh/go-spew 5215b55f46b2b919f50a1df0eaa5886afe4e3b3d
-github.com/docker/engine-api 8924d6900370b4c7e7984be5adc61f50a80d7537
-github.com/docker/go-connections f549a9393d05688dff0992ef3efd8bbe6c628aeb
-github.com/docker/go-units 5d2041e26a699eaca682e2ea41c8f891e1060444
+github.com/davecgh/go-spew 346938d642f2ec3594ed81d874461961cd0faa76
+github.com/docker/docker b89aff1afa1f61993ab2ba18fd62d9375a195f5d
 github.com/eapache/go-resiliency b86b1ec0dd4209a588dc1285cdd471e73525c0b3
-github.com/eapache/queue ded5959c0d4e360646dc9e9908cff48666781367
-github.com/eclipse/paho.mqtt.golang 0f7a459f04f13a41b7ed752d47944528d4bf9a86
-github.com/go-sql-driver/mysql 1fca743146605a172a266e1654e01e5cd5669bee
-github.com/gobwas/glob 49571a1557cd20e6a2410adc6421f85b66c730b5
-github.com/golang/protobuf 552c7b9542c194800fd493123b3798ef0a832032
-github.com/golang/snappy d9eb7a3d35ec988b8585d4a0068e462c27d28380
-github.com/gonuts/go-shellquote e842a11b24c6abfb3dd27af69a17f482e4b483c2
-github.com/gorilla/context 1ea25387ff6f684839d82767c1733ff4d4d15d0a
-github.com/gorilla/mux c9e326e2bdec29039a3761c07bece13133863e1e
+github.com/eapache/go-xerial-snappy bb955e01b9346ac19dc29eb16586c90ded99a98c
+github.com/eapache/queue 44cc805cf13205b55f69e14bcb69867d1ae92f98
+github.com/eclipse/paho.mqtt.golang d4f545eb108a2d19f9b1a735689dbfb719bc21fb
+github.com/go-sql-driver/mysql 2e00b5cd70399450106cec6431c2e2ce3cae5034
+github.com/gobwas/glob bea32b9cd2d6f55753d94a28e959b13f0244797a
+github.com/golang/protobuf 8ee79997227bf9b34611aee7946ae64735e6fd93
+github.com/golang/snappy 7db9049039a047d955fe8c19b83c8ff5abd765c7
+github.com/gorilla/mux 392c28fe23e1c45ddba891b0320b3b5df220beea
 github.com/hailocab/go-hostpool e80d13ce29ede4452c43dea11e79b9bc8a15b478
-github.com/hashicorp/consul 5aa90455ce78d4d41578bafc86305e6e6b28d7d2
-github.com/hpcloud/tail b2940955ab8b26e19d43a43c4da0475dd81bdb56
-github.com/influxdata/config b79f6829346b8d6e78ba73544b1e1038f1f1c9da
-github.com/influxdata/influxdb fc57c0f7c635df3873f3d64f0ed2100ddc94d5ae
-github.com/influxdata/toml af4df43894b16e3fd2b788d01bd27ad0776ef2d0
+github.com/hashicorp/consul 63d2fc68239b996096a1c55a0d4b400ea4c2583f
+github.com/influxdata/tail a395bf99fe07c233f41fba0735fa2b13b58588ea
+github.com/influxdata/toml 5d1d907f22ead1cd47adde17ceec5bda9cacaf8f
 github.com/influxdata/wlog 7c63b0a71ef8300adc255344d275e10e5c3a71ec
-github.com/kardianos/osext 29ae4ffbc9a6fe9fb2bc5029050ce6996ea1d3bc
-github.com/kardianos/service 5e335590050d6d00f3aa270217d288dda1c94d0a
-github.com/klauspost/crc32 19b0b332c9e4516a6370a0456e6182c3b5036720
-github.com/lib/pq e182dc4027e2ded4b19396d638610f2653295f36
-github.com/matttproud/golang_protobuf_extensions d0c3fe89de86839aecf2e0579c40ba3bb336a453
-github.com/miekg/dns cce6c130cdb92c752850880fd285bea1d64439dd
-github.com/mreiferson/go-snappystream 028eae7ab5c4c9e2d1cb4c4ca1e53259bbe7e504
+github.com/jackc/pgx b84338d7d62598f75859b2b146d830b22f1b9ec8
+github.com/kardianos/osext c2c54e542fb797ad986b31721e1baedf214ca413
+github.com/kardianos/service 6d3a0ee7d3425d9d835debc51a0ca1ffa28f4893
+github.com/kballard/go-shellquote d8ec1a69a250a17bb0e419c386eac1f3711dc142
+github.com/klauspost/crc32 cb6bfca970f6908083f26f39a79009d608efd5cd
+github.com/matttproud/golang_protobuf_extensions c12348ce28de40eed0136aa2b644d0ee0650e56c
+github.com/miekg/dns 99f84ae56e75126dd77e5de4fae2ea034a468ca1
 github.com/naoina/go-stringutil 6b638e95a32d0c1131db0e7fe83775cbea4a0d0b
-github.com/nats-io/nats ea8b4fd12ebb823073c0004b9f09ac8748f4f165
-github.com/nats-io/nuid a5152d67cf63cbfb5d992a395458722a45194715
-github.com/nsqio/go-nsq 0b80d6f05e15ca1930e0c5e1d540ed627e299980
-github.com/opencontainers/runc 89ab7f2ccc1e45ddf6485eaa802c35dcf321dfc8
-github.com/prometheus/client_golang 18acf9993a863f4c4b40612e19cdd243e7c86831
+github.com/nats-io/go-nats ea9585611a4ab58a205b9b125ebd74c389a6b898
+github.com/nats-io/nats ea9585611a4ab58a205b9b125ebd74c389a6b898
+github.com/nats-io/nuid 289cccf02c178dc782430d534e3c1f5b72af807f
+github.com/nsqio/go-nsq a53d495e81424aaf7a7665a9d32a97715c40e953
+github.com/pierrec/lz4 5c9560bfa9ace2bf86080bf40d46b34ae44604df
+github.com/pierrec/xxHash 5a004441f897722c627870a981d02b29924215fa
+github.com/prometheus/client_golang c317fb74746eac4fc65fe3909195f4cf67c5562a
 github.com/prometheus/client_model fa8ad6fec33561be4280a8f0514318c79d7f6cb6
-github.com/prometheus/common e8eabff8812b05acf522b45fdcd725a785188e37
-github.com/prometheus/procfs 406e5b7bfd8201a36e2bb5f7bdae0b03380c2ce8
-github.com/samuel/go-zookeeper 218e9c81c0dd8b3b18172b2bbfad92cc7d6db55f
-github.com/shirou/gopsutil 4d0c402af66c78735c5ccf820dc2ca7de5e4ff08
-github.com/soniah/gosnmp 3fe3beb30fa9700988893c56a63b1df8e1b68c26
-github.com/streadway/amqp b4f3ceab0337f013208d31348b578d83c0064744
-github.com/stretchr/testify 1f4a1643a57e798696635ea4c126e9127adb7d3c
-github.com/vjeantet/grok 83bfdfdfd1a8146795b28e547a8e3c8b28a466c2
-github.com/wvanbergen/kafka 46f9a1cf3f670edec492029fadded9c2d9e18866
-github.com/wvanbergen/kazoo-go 0f768712ae6f76454f987c3356177e138df258f8
-github.com/yuin/gopher-lua bf3808abd44b1e55143a2d7f08571aaa80db1808
+github.com/prometheus/common dd2f054febf4a6c00f2343686efb775948a8bff4
+github.com/prometheus/procfs 1878d9fbb537119d24b21ca07effd591627cd160
+github.com/rcrowley/go-metrics 1f30fe9094a513ce4c700b9a54458bbb0c96996c
+github.com/samuel/go-zookeeper 1d7be4effb13d2d908342d349d71a284a7542693
+github.com/satori/go.uuid 5bf94b69c6b68ee1b541973bb8e1144db23a194b
+github.com/shirou/gopsutil 9a4a9167ad3b4355dbf1c2c7a0f5f0d3fb1e9ab9
+github.com/soniah/gosnmp 5ad50dc75ab389f8a1c9f8a67d3a1cd85f67ed15
+github.com/streadway/amqp 63795daa9a446c920826655f26ba31c81c860fd6
+github.com/stretchr/testify 4d4bfba8f1d1027c4fdbe371823030df51419987
+github.com/vjeantet/grok d73e972b60935c7fec0b4ffbc904ed39ecaf7efe
+github.com/wvanbergen/kafka bc265fedb9ff5b5c5d3c0fdcef4a819b3523d3ee
+github.com/wvanbergen/kazoo-go 968957352185472eacb69215fa3dbfcfdbac1096
+github.com/yuin/gopher-lua 66c871e454fcf10251c61bf8eff02d0978cae75a
 github.com/zensqlmonitor/go-mssqldb ffe5510c6fa5e15e6d983210ab501c815b56b363
-golang.org/x/crypto c197bcf24cde29d3f73c7b4ac6fd41f4384e8af6
-golang.org/x/net 6acef71eb69611914f7a30939ea9f6e194c78172
-golang.org/x/text a71fd10341b064c10f4a81ceac72bcf70f26ea34
-gopkg.in/dancannon/gorethink.v1 7d1af5be49cb5ecc7b177bf387d232050299d6ef
-gopkg.in/fatih/pool.v2 cba550ebf9bce999a02e963296d4bc7a486cb715
-gopkg.in/mgo.v2 d90005c5262a3463800497ea5a89aed5fe22c886
-gopkg.in/yaml.v2 a83829b6f1293c91addabc89d0571c246397bbf4
+golang.org/x/crypto dc137beb6cce2043eb6b5f223ab8bf51c32459f4
+golang.org/x/net f2499483f923065a842d38eb4c7f1927e6fc6e6d
+golang.org/x/text 506f9d5c962f284575e88337e7d9296d27e729d3
+gopkg.in/dancannon/gorethink.v1 edc7a6a68e2d8015f5ffe1b2560eed989f8a45be
+gopkg.in/fatih/pool.v2 6e328e67893eb46323ad06f0e92cb9536babbabc
+gopkg.in/mgo.v2 3f83fa5005286a7fe593b055f0d7771a7dce4655
+gopkg.in/olivere/elastic.v5 ee3ebceab960cf68ab9a89ee6d78c031ef5b4a4e
+gopkg.in/yaml.v2 4c78c975fe7c825c6d1466c42be594d1d6f3aba6

Godeps_windows

@@ -1,7 +1,6 @@
 github.com/Microsoft/go-winio ce2922f643c8fd76b46cadc7f404a06282678b34
 github.com/StackExchange/wmi f3e2bae1e0cb5aef83e319133eabfee30013a4a5
 github.com/go-ole/go-ole be49f7c07711fcb603cff39e1de7c67926dc0ba7
-github.com/lxn/win 950a0e81e7678e63d8e6cd32412bdecb325ccd88
 github.com/shirou/w32 3c9377fc6748f222729a8270fe2775d149a249ad
 golang.org/x/sys a646d33e2ee3172a661fc09bca23bb4889a41bc8
 github.com/go-ini/ini 9144852efba7c4daf409943ee90767da62d55438

Makefile

@@ -1,6 +1,6 @@
 VERSION := $(shell sh -c 'git describe --always --tags')
 BRANCH := $(shell sh -c 'git rev-parse --abbrev-ref HEAD')
-COMMIT := $(shell sh -c 'git rev-parse HEAD')
+COMMIT := $(shell sh -c 'git rev-parse --short HEAD')
 ifdef GOBIN
 PATH := $(GOBIN):$(PATH)
 else
@@ -51,6 +51,7 @@ docker-run:
 		-e ADVERTISED_PORT=9092 \
 		-p "2181:2181" -p "9092:9092" \
 		-d spotify/kafka
+	docker run --name elasticsearch -p "9200:9200" -p "9300:9300" -d elasticsearch:5
 	docker run --name mysql -p "3306:3306" -e MYSQL_ALLOW_EMPTY_PASSWORD=yes -d mysql
 	docker run --name memcached -p "11211:11211" -d memcached
 	docker run --name postgres -p "5432:5432" -d postgres
@@ -58,7 +59,7 @@ docker-run:
 	docker run --name redis -p "6379:6379" -d redis
 	docker run --name nsq -p "4150:4150" -d nsqio/nsq /nsqd
 	docker run --name mqtt -p "1883:1883" -d ncarlier/mqtt
-	docker run --name riemann -p "5555:5555" -d blalor/riemann
+	docker run --name riemann -p "5555:5555" -d stealthly/docker-riemann
 	docker run --name nats -p "4222:4222" -d nats
 
 # Run docker containers necessary for CircleCI unit tests
@@ -69,15 +70,16 @@ docker-run-circle:
 		-e ADVERTISED_PORT=9092 \
 		-p "2181:2181" -p "9092:9092" \
 		-d spotify/kafka
+	docker run --name elasticsearch -p "9200:9200" -p "9300:9300" -d elasticsearch:5
 	docker run --name nsq -p "4150:4150" -d nsqio/nsq /nsqd
 	docker run --name mqtt -p "1883:1883" -d ncarlier/mqtt
-	docker run --name riemann -p "5555:5555" -d blalor/riemann
+	docker run --name riemann -p "5555:5555" -d stealthly/docker-riemann
 	docker run --name nats -p "4222:4222" -d nats
 
 # Kill all docker containers, ignore errors
 docker-kill:
-	-docker kill nsq aerospike redis rabbitmq postgres memcached mysql kafka mqtt riemann nats
-	-docker rm nsq aerospike redis rabbitmq postgres memcached mysql kafka mqtt riemann nats
+	-docker kill nsq aerospike redis rabbitmq postgres memcached mysql kafka mqtt riemann nats elasticsearch
+	-docker rm nsq aerospike redis rabbitmq postgres memcached mysql kafka mqtt riemann nats elasticsearch
 
 # Run full unit tests using docker containers (includes setup and teardown)
 test: vet docker-kill docker-run

README.md

@@ -1,15 +1,23 @@
 # Telegraf [![Circle CI](https://circleci.com/gh/influxdata/telegraf.svg?style=svg)](https://circleci.com/gh/influxdata/telegraf) [![Docker pulls](https://img.shields.io/docker/pulls/library/telegraf.svg)](https://hub.docker.com/_/telegraf/)
 
-Telegraf is an agent written in Go for collecting metrics from the system it's
-running on, or from other services, and writing them into InfluxDB or other
-[outputs](https://github.com/influxdata/telegraf#supported-output-plugins).
+Telegraf is an agent written in Go for collecting, processing, aggregating,
+and writing metrics.
 
 Design goals are to have a minimal memory footprint with a plugin system so
 that developers in the community can easily add support for collecting metrics
 from well known services (like Hadoop, Postgres, or Redis) and third party
 APIs (like Mailchimp, AWS CloudWatch, or Google Analytics).
 
-New input and output plugins are designed to be easy to contribute,
+Telegraf is plugin-driven and has the concept of 4 distinct plugins:
+
+1. [Input Plugins](#input-plugins) collect metrics from the system, services, or 3rd party APIs
+2. [Processor Plugins](#processor-plugins) transform, decorate, and/or filter metrics
+3. [Aggregator Plugins](#aggregator-plugins) create aggregate metrics (e.g. mean, min, max, quantiles, etc.)
+4. [Output Plugins](#output-plugins) write metrics to various destinations
+
+For more information on Processor and Aggregator plugins please [read this](./docs/AGGREGATORS_AND_PROCESSORS.md).
+
+New plugins are designed to be easy to contribute,
 we'll eagerly accept pull
 requests and will manage the set of plugins that Telegraf supports.
 See the [contributing guide](CONTRIBUTING.md) for instructions on writing
@@ -17,65 +25,25 @@ new plugins.
 
 ## Installation:
 
-### Linux deb and rpm Packages:
+You can either download the binaries directly from the
+[downloads](https://www.influxdata.com/downloads) page.
 
-Latest:
-* https://dl.influxdata.com/telegraf/releases/telegraf_1.0.1_amd64.deb
-* https://dl.influxdata.com/telegraf/releases/telegraf-1.0.1.x86_64.rpm
-
-Latest (arm):
-* https://dl.influxdata.com/telegraf/releases/telegraf_1.0.1_armhf.deb
-* https://dl.influxdata.com/telegraf/releases/telegraf-1.0.1.armhf.rpm
-
-##### Package Instructions:
-
-* Telegraf binary is installed in `/usr/bin/telegraf`
-* Telegraf daemon configuration file is in `/etc/telegraf/telegraf.conf`
-* On sysv systems, the telegraf daemon can be controlled via
-`service telegraf [action]`
-* On systemd systems (such as Ubuntu 15+), the telegraf daemon can be
-controlled via `systemctl [action] telegraf`
-
-### yum/apt Repositories:
-
-There is a yum/apt repo available for the whole InfluxData stack, see
-[here](https://docs.influxdata.com/influxdb/v0.10/introduction/installation/#installation)
-for instructions on setting up the repo. Once it is configured, you will be able
-to use this repo to install & update telegraf.
-
-### Linux tarballs:
-
-Latest:
-* https://dl.influxdata.com/telegraf/releases/telegraf-1.0.1_linux_amd64.tar.gz
-* https://dl.influxdata.com/telegraf/releases/telegraf-1.0.1_linux_i386.tar.gz
-* https://dl.influxdata.com/telegraf/releases/telegraf-1.0.1_linux_armhf.tar.gz
+A few alternate installs are available here as well:
 
 ### FreeBSD tarball:
 
 Latest:
-* https://dl.influxdata.com/telegraf/releases/telegraf-1.0.1_freebsd_amd64.tar.gz
+* https://dl.influxdata.com/telegraf/releases/telegraf-VERSION_freebsd_amd64.tar.gz
 
 ### Ansible Role:
 
 Ansible role: https://github.com/rossmcdonald/telegraf
 
-### OSX via Homebrew:
-
-```
-brew update
-brew install telegraf
-```
-
-### Windows Binaries (EXPERIMENTAL)
-
-Latest:
-* https://dl.influxdata.com/telegraf/releases/telegraf-1.0.1_windows_amd64.zip
-
 ### From Source:
 
 Telegraf manages dependencies via [gdm](https://github.com/sparrc/gdm),
 which gets installed via the Makefile
-if you don't have it already. You also must build with golang version 1.5+.
+if you don't have it already. You also must build with golang version 1.8+.
 
 1. [Install Go](https://golang.org/doc/install)
 2. [Setup your GOPATH](https://golang.org/doc/code.html#GOPATH)
@@ -91,31 +59,31 @@ See usage with:
 telegraf --help
 ```
 
-### Generate a telegraf config file:
+#### Generate a telegraf config file:
 
 ```
 telegraf config > telegraf.conf
 ```
 
-### Generate config with only cpu input & influxdb output plugins defined
+#### Generate config with only cpu input & influxdb output plugins defined
 
 ```
 telegraf --input-filter cpu --output-filter influxdb config
 ```
 
-### Run a single telegraf collection, outputing metrics to stdout
+#### Run a single telegraf collection, outputing metrics to stdout
 
 ```
 telegraf --config telegraf.conf -test
 ```
 
-### Run telegraf with all plugins defined in config file
+#### Run telegraf with all plugins defined in config file
 
 ```
 telegraf --config telegraf.conf
 ```
 
-### Run telegraf, enabling the cpu & memory input, and influxdb output plugins
+#### Run telegraf, enabling the cpu & memory input, and influxdb output plugins
 
 ```
 telegraf --config telegraf.conf -input-filter cpu:mem -output-filter influxdb
@@ -127,77 +95,78 @@ telegraf --config telegraf.conf -input-filter cpu:mem -output-filter influxdb
 See the [configuration guide](docs/CONFIGURATION.md) for a rundown of the more advanced
 configuration options.
 
-## Supported Input Plugins
+## Input Plugins
 
-Telegraf currently has support for collecting metrics from many sources. For
-more information on each, please look at the directory of the same name in
-`plugins/inputs`.
-
-Currently implemented sources:
-
-* [aws cloudwatch](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/cloudwatch)
-* [aerospike](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/aerospike)
-* [apache](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/apache)
-* [bcache](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/bcache)
-* [cassandra](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/cassandra)
-* [ceph](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/ceph)
-* [chrony](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/chrony)
-* [consul](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/consul)
-* [conntrack](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/conntrack)
-* [couchbase](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/couchbase)
-* [couchdb](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/couchdb)
-* [disque](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/disque)
-* [dns query time](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/dns_query)
-* [docker](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/docker)
-* [dovecot](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/dovecot)
-* [elasticsearch](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/elasticsearch)
-* [exec](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/exec) (generic executable plugin, support JSON, influx, graphite and nagios)
-* [filestat](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/filestat)
-* [haproxy](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/haproxy)
-* [hddtemp](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/hddtemp)
-* [http_response](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/http_response)
-* [httpjson](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/httpjson) (generic JSON-emitting http service plugin)
-* [influxdb](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/influxdb)
-* [ipmi_sensor](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/ipmi_sensor)
-* [iptables](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/iptables)
-* [jolokia](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/jolokia)
-* [leofs](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/leofs)
-* [lustre2](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/lustre2)
-* [mailchimp](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/mailchimp)
-* [memcached](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/memcached)
-* [mesos](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/mesos)
-* [mongodb](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/mongodb)
-* [mysql](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/mysql)
-* [net_response](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/net_response)
-* [nginx](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/nginx)
-* [nsq](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/nsq)
-* [nstat](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/nstat)
-* [ntpq](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/ntpq)
-* [phpfpm](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/phpfpm)
-* [phusion passenger](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/passenger)
-* [ping](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/ping)
-* [postgresql](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/postgresql)
-* [postgresql_extensible](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/postgresql_extensible)
-* [powerdns](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/powerdns)
-* [procstat](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/procstat)
-* [prometheus](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/prometheus)
-* [puppetagent](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/puppetagent)
-* [rabbitmq](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/rabbitmq)
-* [raindrops](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/raindrops)
-* [redis](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/redis)
-* [rethinkdb](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/rethinkdb)
-* [riak](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/riak)
-* [sensors](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/sensors)
-* [snmp](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/snmp)
-* [snmp_legacy](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/snmp_legacy)
-* [sql server](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/sqlserver) (microsoft)
-* [twemproxy](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/twemproxy)
-* [varnish](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/varnish)
-* [zfs](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/zfs)
-* [zookeeper](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/zookeeper)
-* [win_perf_counters ](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/win_perf_counters) (windows performance counters)
-* [sysstat](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/sysstat)
-* [system](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/system)
+* [aerospike](./plugins/inputs/aerospike)
+* [amqp_consumer](./plugins/inputs/amqp_consumer) (rabbitmq)
+* [apache](./plugins/inputs/apache)
+* [aws cloudwatch](./plugins/inputs/cloudwatch)
+* [bcache](./plugins/inputs/bcache)
+* [cassandra](./plugins/inputs/cassandra)
+* [ceph](./plugins/inputs/ceph)
+* [cgroup](./plugins/inputs/cgroup)
+* [chrony](./plugins/inputs/chrony)
+* [consul](./plugins/inputs/consul)
+* [conntrack](./plugins/inputs/conntrack)
+* [couchbase](./plugins/inputs/couchbase)
+* [couchdb](./plugins/inputs/couchdb)
+* [disque](./plugins/inputs/disque)
+* [dmcache](./plugins/inputs/dmcache)
+* [dns query time](./plugins/inputs/dns_query)
+* [docker](./plugins/inputs/docker)
+* [dovecot](./plugins/inputs/dovecot)
+* [elasticsearch](./plugins/inputs/elasticsearch)
+* [exec](./plugins/inputs/exec) (generic executable plugin, support JSON, influx, graphite and nagios)
+* [filestat](./plugins/inputs/filestat)
+* [haproxy](./plugins/inputs/haproxy)
+* [hddtemp](./plugins/inputs/hddtemp)
+* [http_response](./plugins/inputs/http_response)
+* [httpjson](./plugins/inputs/httpjson) (generic JSON-emitting http service plugin)
+* [internal](./plugins/inputs/internal)
+* [influxdb](./plugins/inputs/influxdb)
+* [interrupts](./plugins/inputs/interrupts)
+* [ipmi_sensor](./plugins/inputs/ipmi_sensor)
+* [iptables](./plugins/inputs/iptables)
+* [jolokia](./plugins/inputs/jolokia)
+* [kapacitor](./plugins/inputs/kapacitor)
+* [kubernetes](./plugins/inputs/kubernetes)
+* [leofs](./plugins/inputs/leofs)
+* [lustre2](./plugins/inputs/lustre2)
+* [mailchimp](./plugins/inputs/mailchimp)
+* [memcached](./plugins/inputs/memcached)
+* [mesos](./plugins/inputs/mesos)
+* [mongodb](./plugins/inputs/mongodb)
+* [mysql](./plugins/inputs/mysql)
+* [net_response](./plugins/inputs/net_response)
+* [nginx](./plugins/inputs/nginx)
+* [nsq](./plugins/inputs/nsq)
+* [nstat](./plugins/inputs/nstat)
+* [ntpq](./plugins/inputs/ntpq)
+* [phpfpm](./plugins/inputs/phpfpm)
+* [phusion passenger](./plugins/inputs/passenger)
+* [ping](./plugins/inputs/ping)
+* [postgresql](./plugins/inputs/postgresql)
+* [postgresql_extensible](./plugins/inputs/postgresql_extensible)
+* [powerdns](./plugins/inputs/powerdns)
+* [procstat](./plugins/inputs/procstat)
+* [prometheus](./plugins/inputs/prometheus)
+* [puppetagent](./plugins/inputs/puppetagent)
+* [rabbitmq](./plugins/inputs/rabbitmq)
+* [raindrops](./plugins/inputs/raindrops)
+* [redis](./plugins/inputs/redis)
+* [rethinkdb](./plugins/inputs/rethinkdb)
+* [riak](./plugins/inputs/riak)
+* [sensors](./plugins/inputs/sensors)
+* [snmp](./plugins/inputs/snmp)
+* [snmp_legacy](./plugins/inputs/snmp_legacy)
+* [sql server](./plugins/inputs/sqlserver) (microsoft)
+* [twemproxy](./plugins/inputs/twemproxy)
+* [varnish](./plugins/inputs/varnish)
+* [zfs](./plugins/inputs/zfs)
+* [zookeeper](./plugins/inputs/zookeeper)
+* [win_perf_counters ](./plugins/inputs/win_perf_counters) (windows performance counters)
+* [sysstat](./plugins/inputs/sysstat)
+* [system](./plugins/inputs/system)
     * cpu
     * mem
     * net
@@ -208,48 +177,72 @@ Currently implemented sources:
     * processes
     * kernel (/proc/stat)
     * kernel (/proc/vmstat)
+    * linux_sysctl_fs (/proc/sys/fs)
 
 Telegraf can also collect metrics via the following service plugins:
 
-* [http_listener](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/http_listener)
-* [kafka_consumer](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/kafka_consumer)
-* [mqtt_consumer](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/mqtt_consumer)
-* [nats_consumer](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/nats_consumer)
-* [nsq_consumer](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/nsq_consumer)
-* [logparser](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/logparser)
-* [statsd](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/statsd)
-* [tail](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/tail)
-* [tcp_listener](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/tcp_listener)
-* [udp_listener](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/udp_listener)
-* [webhooks](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/webhooks)
-  * [filestack](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/webhooks/filestack)
-  * [github](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/webhooks/github)
-  * [mandrill](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/webhooks/mandrill)
-  * [rollbar](https://github.com/influxdata/telegraf/tree/master/plugins/inputs/webhooks/rollbar)
+* [http_listener](./plugins/inputs/http_listener)
+* [kafka_consumer](./plugins/inputs/kafka_consumer)
+* [mqtt_consumer](./plugins/inputs/mqtt_consumer)
+* [nats_consumer](./plugins/inputs/nats_consumer)
+* [nsq_consumer](./plugins/inputs/nsq_consumer)
+* [logparser](./plugins/inputs/logparser)
+* [statsd](./plugins/inputs/statsd)
+* [socket_listener](./plugins/inputs/socket_listener)
+* [tail](./plugins/inputs/tail)
+* [tcp_listener](./plugins/inputs/socket_listener)
+* [udp_listener](./plugins/inputs/socket_listener)
+* [webhooks](./plugins/inputs/webhooks)
+  * [filestack](./plugins/inputs/webhooks/filestack)
+  * [github](./plugins/inputs/webhooks/github)
+  * [mandrill](./plugins/inputs/webhooks/mandrill)
+  * [rollbar](./plugins/inputs/webhooks/rollbar)
+  * [papertrail](./plugins/inputs/webhooks/papertrail)
 
-We'll be adding support for many more over the coming months. Read on if you
-want to add support for another service or third-party API.
+Telegraf is able to parse the following input data formats into metrics, these
+formats may be used with input plugins supporting the `data_format` option:
 
-## Supported Output Plugins
+* [InfluxDB Line Protocol](./docs/DATA_FORMATS_INPUT.md#influx)
+* [JSON](./docs/DATA_FORMATS_INPUT.md#json)
+* [Graphite](./docs/DATA_FORMATS_INPUT.md#graphite)
+* [Value](./docs/DATA_FORMATS_INPUT.md#value)
+* [Nagios](./docs/DATA_FORMATS_INPUT.md#nagios)
+* [Collectd](./docs/DATA_FORMATS_INPUT.md#collectd)
 
-* [influxdb](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/influxdb)
-* [amon](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/amon)
-* [amqp](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/amqp)
-* [aws kinesis](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/kinesis)
-* [aws cloudwatch](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/cloudwatch)
-* [datadog](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/datadog)
-* [file](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/file)
-* [graphite](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/graphite)
-* [graylog](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/graylog)
-* [instrumental](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/instrumental)
-* [kafka](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/kafka)
-* [librato](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/librato)
-* [mqtt](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/mqtt)
-* [nats](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/nats)
-* [nsq](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/nsq)
-* [opentsdb](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/opentsdb)
-* [prometheus](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/prometheus_client)
-* [riemann](https://github.com/influxdata/telegraf/tree/master/plugins/outputs/riemann)
+## Processor Plugins
+
+* [printer](./plugins/processors/printer)
+
+## Aggregator Plugins
+
+* [minmax](./plugins/aggregators/minmax)
+
+## Output Plugins
+
+* [influxdb](./plugins/outputs/influxdb)
+* [amon](./plugins/outputs/amon)
+* [amqp](./plugins/outputs/amqp) (rabbitmq)
+* [aws kinesis](./plugins/outputs/kinesis)
+* [aws cloudwatch](./plugins/outputs/cloudwatch)
+* [datadog](./plugins/outputs/datadog)
+* [discard](./plugins/outputs/discard)
+* [elasticsearch](./plugins/outputs/elasticsearch)
+* [file](./plugins/outputs/file)
+* [graphite](./plugins/outputs/graphite)
+* [graylog](./plugins/outputs/graylog)
+* [instrumental](./plugins/outputs/instrumental)
+* [kafka](./plugins/outputs/kafka)
+* [librato](./plugins/outputs/librato)
+* [mqtt](./plugins/outputs/mqtt)
+* [nats](./plugins/outputs/nats)
+* [nsq](./plugins/outputs/nsq)
+* [opentsdb](./plugins/outputs/opentsdb)
+* [prometheus](./plugins/outputs/prometheus_client)
+* [riemann](./plugins/outputs/riemann)
+* [riemann_legacy](./plugins/outputs/riemann_legacy)
+* [socket_writer](./plugins/outputs/socket_writer)
+* [tcp](./plugins/outputs/socket_writer)
+* [udp](./plugins/outputs/socket_writer)
 
 ## Contributing
 

agent/accumulator.go

@@ -2,10 +2,14 @@ package agent
 
 import (
 	"log"
-	"sync/atomic"
 	"time"
 
 	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/selfstat"
+)
+
+var (
+	NErrors = selfstat.Register("agent", "gather_errors", map[string]string{})
 )
 
 type MetricMaker interface {
@@ -37,8 +41,6 @@ type accumulator struct {
 	maker MetricMaker
 
 	precision time.Duration
-
-	errCount uint64
 }
 
 func (ac *accumulator) AddFields(
@@ -80,7 +82,7 @@ func (ac *accumulator) AddError(err error) {
 	if err == nil {
 		return
 	}
-	atomic.AddUint64(&ac.errCount, 1)
+	NErrors.Incr(1)
 	//TODO suppress/throttle consecutive duplicate errors?
 	log.Printf("E! Error in plugin [%s]: %s", ac.maker.Name(), err)
 }

agent/accumulator_test.go

@@ -9,6 +9,7 @@ import (
 	"time"
 
 	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/metric"
 
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -41,7 +42,7 @@ func TestAdd(t *testing.T) {
 	testm = <-metrics
 	actual = testm.String()
 	assert.Equal(t,
-		fmt.Sprintf("acctest,acc=test value=101 %d", now.UnixNano()),
+		fmt.Sprintf("acctest,acc=test value=101 %d\n", now.UnixNano()),
 		actual)
 }
 
@@ -69,7 +70,7 @@ func TestAddFields(t *testing.T) {
 	testm = <-metrics
 	actual = testm.String()
 	assert.Equal(t,
-		fmt.Sprintf("acctest,acc=test usage=99 %d", now.UnixNano()),
+		fmt.Sprintf("acctest,acc=test usage=99 %d\n", now.UnixNano()),
 		actual)
 }
 
@@ -87,7 +88,7 @@ func TestAccAddError(t *testing.T) {
 	a.AddError(fmt.Errorf("baz"))
 
 	errs := bytes.Split(errBuf.Bytes(), []byte{'\n'})
-	assert.EqualValues(t, 3, a.errCount)
+	assert.EqualValues(t, int64(3), NErrors.Get())
 	require.Len(t, errs, 4) // 4 because of trailing newline
 	assert.Contains(t, string(errs[0]), "TestPlugin")
 	assert.Contains(t, string(errs[0]), "foo")
@@ -125,7 +126,7 @@ func TestAddNoIntervalWithPrecision(t *testing.T) {
 	testm = <-a.metrics
 	actual = testm.String()
 	assert.Equal(t,
-		fmt.Sprintf("acctest,acc=test value=101 %d", int64(1139572800000000000)),
+		fmt.Sprintf("acctest,acc=test value=101 %d\n", int64(1139572800000000000)),
 		actual)
 }
 
@@ -157,7 +158,7 @@ func TestAddDisablePrecision(t *testing.T) {
 	testm = <-a.metrics
 	actual = testm.String()
 	assert.Equal(t,
-		fmt.Sprintf("acctest,acc=test value=101 %d", int64(1139572800082912748)),
+		fmt.Sprintf("acctest,acc=test value=101 %d\n", int64(1139572800082912748)),
 		actual)
 }
 
@@ -189,7 +190,7 @@ func TestAddNoPrecisionWithInterval(t *testing.T) {
 	testm = <-a.metrics
 	actual = testm.String()
 	assert.Equal(t,
-		fmt.Sprintf("acctest,acc=test value=101 %d", int64(1139572800000000000)),
+		fmt.Sprintf("acctest,acc=test value=101 %d\n", int64(1139572800000000000)),
 		actual)
 }
 
@@ -206,7 +207,7 @@ func TestDifferentPrecisions(t *testing.T) {
 	testm := <-a.metrics
 	actual := testm.String()
 	assert.Equal(t,
-		fmt.Sprintf("acctest,acc=test value=101 %d", int64(1139572800000000000)),
+		fmt.Sprintf("acctest,acc=test value=101 %d\n", int64(1139572800000000000)),
 		actual)
 
 	a.SetPrecision(0, time.Millisecond)
@@ -216,7 +217,7 @@ func TestDifferentPrecisions(t *testing.T) {
 	testm = <-a.metrics
 	actual = testm.String()
 	assert.Equal(t,
-		fmt.Sprintf("acctest,acc=test value=101 %d", int64(1139572800083000000)),
+		fmt.Sprintf("acctest,acc=test value=101 %d\n", int64(1139572800083000000)),
 		actual)
 
 	a.SetPrecision(0, time.Microsecond)
@@ -226,7 +227,7 @@ func TestDifferentPrecisions(t *testing.T) {
 	testm = <-a.metrics
 	actual = testm.String()
 	assert.Equal(t,
-		fmt.Sprintf("acctest,acc=test value=101 %d", int64(1139572800082913000)),
+		fmt.Sprintf("acctest,acc=test value=101 %d\n", int64(1139572800082913000)),
 		actual)
 
 	a.SetPrecision(0, time.Nanosecond)
@@ -236,7 +237,7 @@ func TestDifferentPrecisions(t *testing.T) {
 	testm = <-a.metrics
 	actual = testm.String()
 	assert.Equal(t,
-		fmt.Sprintf("acctest,acc=test value=101 %d", int64(1139572800082912748)),
+		fmt.Sprintf("acctest,acc=test value=101 %d\n", int64(1139572800082912748)),
 		actual)
 }
 
@@ -269,7 +270,7 @@ func TestAddGauge(t *testing.T) {
 	testm = <-metrics
 	actual = testm.String()
 	assert.Equal(t,
-		fmt.Sprintf("acctest,acc=test value=101 %d", now.UnixNano()),
+		fmt.Sprintf("acctest,acc=test value=101 %d\n", now.UnixNano()),
 		actual)
 	assert.Equal(t, testm.Type(), telegraf.Gauge)
 }
@@ -303,7 +304,7 @@ func TestAddCounter(t *testing.T) {
 	testm = <-metrics
 	actual = testm.String()
 	assert.Equal(t,
-		fmt.Sprintf("acctest,acc=test value=101 %d", now.UnixNano()),
+		fmt.Sprintf("acctest,acc=test value=101 %d\n", now.UnixNano()),
 		actual)
 	assert.Equal(t, testm.Type(), telegraf.Counter)
 }
@@ -323,15 +324,15 @@ func (tm *TestMetricMaker) MakeMetric(
 ) telegraf.Metric {
 	switch mType {
 	case telegraf.Untyped:
-		if m, err := telegraf.NewMetric(measurement, tags, fields, t); err == nil {
+		if m, err := metric.New(measurement, tags, fields, t); err == nil {
 			return m
 		}
 	case telegraf.Counter:
-		if m, err := telegraf.NewCounterMetric(measurement, tags, fields, t); err == nil {
+		if m, err := metric.New(measurement, tags, fields, t, telegraf.Counter); err == nil {
 			return m
 		}
 	case telegraf.Gauge:
-		if m, err := telegraf.NewGaugeMetric(measurement, tags, fields, t); err == nil {
+		if m, err := metric.New(measurement, tags, fields, t, telegraf.Gauge); err == nil {
 			return m
 		}
 	}

agent/agent.go

@@ -12,6 +12,7 @@ import (
 	"github.com/influxdata/telegraf/internal"
 	"github.com/influxdata/telegraf/internal/config"
 	"github.com/influxdata/telegraf/internal/models"
+	"github.com/influxdata/telegraf/selfstat"
 )
 
 // Agent runs telegraf and collects data based on the given config
@@ -44,8 +45,6 @@ func NewAgent(config *config.Config) (*Agent, error) {
 // Connect connects to all configured outputs
 func (a *Agent) Connect() error {
 	for _, o := range a.Config.Outputs {
-		o.Quiet = a.Config.Agent.Quiet
-
 		switch ot := o.Output.(type) {
 		case telegraf.ServiceOutput:
 			if err := ot.Start(); err != nil {
@@ -106,24 +105,26 @@ func (a *Agent) gatherer(
 ) {
 	defer panicRecover(input)
 
+	GatherTime := selfstat.RegisterTiming("gather",
+		"gather_time_ns",
+		map[string]string{"input": input.Config.Name},
+	)
+
+	acc := NewAccumulator(input, metricC)
+	acc.SetPrecision(a.Config.Agent.Precision.Duration,
+		a.Config.Agent.Interval.Duration)
+
 	ticker := time.NewTicker(interval)
 	defer ticker.Stop()
 
 	for {
-		acc := NewAccumulator(input, metricC)
-		acc.SetPrecision(a.Config.Agent.Precision.Duration,
-			a.Config.Agent.Interval.Duration)
-		input.SetDebug(a.Config.Agent.Debug)
-		input.SetDefaultTags(a.Config.Tags)
-
 		internal.RandomSleep(a.Config.Agent.CollectionJitter.Duration, shutdown)
 
 		start := time.Now()
 		gatherWithTimeout(shutdown, input, acc, interval)
 		elapsed := time.Since(start)
 
-		log.Printf("D! Input [%s] gathered metrics, (%s interval) in %s\n",
-			input.Name(), interval, elapsed)
+		GatherTime.Incr(elapsed.Nanoseconds())
 
 		select {
 		case <-shutdown:
@@ -156,13 +157,13 @@ func gatherWithTimeout(
 		select {
 		case err := <-done:
 			if err != nil {
-				log.Printf("E! ERROR in input [%s]: %s", input.Name(), err)
+				acc.AddError(err)
 			}
 			return
 		case <-ticker.C:
-			log.Printf("E! ERROR: input [%s] took longer to collect than "+
-				"collection interval (%s)",
-				input.Name(), timeout)
+			err := fmt.Errorf("took longer to collect than collection interval (%s)",
+				timeout)
+			acc.AddError(err)
 			continue
 		case <-shutdown:
 			return
@@ -190,6 +191,12 @@ func (a *Agent) Test() error {
 	}()
 
 	for _, input := range a.Config.Inputs {
+		if _, ok := input.Input.(telegraf.ServiceInput); ok {
+			fmt.Printf("\nWARNING: skipping plugin [[%s]]: service inputs not supported in --test mode\n",
+				input.Name())
+			continue
+		}
+
 		acc := NewAccumulator(input, metricC)
 		acc.SetPrecision(a.Config.Agent.Precision.Duration,
 			a.Config.Agent.Interval.Duration)
@@ -204,14 +211,11 @@ func (a *Agent) Test() error {
 		if err := input.Input.Gather(acc); err != nil {
 			return err
 		}
-		if acc.errCount > 0 {
-			return fmt.Errorf("Errors encountered during processing")
-		}
 
 		// Special instructions for some inputs. cpu, for example, needs to be
 		// run twice in order to return cpu usage percentages.
 		switch input.Name() {
-		case "cpu", "mongodb", "procstat":
+		case "inputs.cpu", "inputs.mongodb", "inputs.procstat":
 			time.Sleep(500 * time.Millisecond)
 			fmt.Printf("* Plugin: %s, Collection 2\n", input.Name())
 			if err := input.Input.Gather(acc); err != nil {
@@ -269,7 +273,7 @@ func (a *Agent) flusher(shutdown chan struct{}, metricC chan telegraf.Metric) er
 				var dropOriginal bool
 				if !m.IsAggregate() {
 					for _, agg := range a.Config.Aggregators {
-						if ok := agg.Add(copyMetric(m)); ok {
+						if ok := agg.Add(m.Copy()); ok {
 							dropOriginal = true
 						}
 					}
@@ -279,7 +283,7 @@ func (a *Agent) flusher(shutdown chan struct{}, metricC chan telegraf.Metric) er
 						if i == len(a.Config.Outputs)-1 {
 							o.AddMetric(m)
 						} else {
-							o.AddMetric(copyMetric(m))
+							o.AddMetric(m.Copy())
 						}
 					}
 				}
@@ -288,6 +292,7 @@ func (a *Agent) flusher(shutdown chan struct{}, metricC chan telegraf.Metric) er
 	}()
 
 	ticker := time.NewTicker(a.Config.Agent.FlushInterval.Duration)
+	semaphore := make(chan struct{}, 1)
 	for {
 		select {
 		case <-shutdown:
@@ -297,8 +302,18 @@ func (a *Agent) flusher(shutdown chan struct{}, metricC chan telegraf.Metric) er
 			a.flush()
 			return nil
 		case <-ticker.C:
-			internal.RandomSleep(a.Config.Agent.FlushJitter.Duration, shutdown)
-			a.flush()
+			go func() {
+				select {
+				case semaphore <- struct{}{}:
+					internal.RandomSleep(a.Config.Agent.FlushJitter.Duration, shutdown)
+					a.flush()
+					<-semaphore
+				default:
+					// skipping this flush because one is already happening
+					log.Println("W! Skipping a scheduled flush because there is" +
+						" already a flush ongoing.")
+				}
+			}()
 		case metric := <-metricC:
 			// NOTE potential bottleneck here as we put each metric through the
 			// processors serially.
@@ -327,13 +342,13 @@ func (a *Agent) Run(shutdown chan struct{}) error {
 
 	// Start all ServicePlugins
 	for _, input := range a.Config.Inputs {
+		input.SetDefaultTags(a.Config.Tags)
 		switch p := input.Input.(type) {
 		case telegraf.ServiceInput:
 			acc := NewAccumulator(input, metricC)
 			// Service input plugins should set their own precision of their
 			// metrics.
 			acc.SetPrecision(time.Nanosecond, 0)
-			input.SetDefaultTags(a.Config.Tags)
 			if err := p.Start(acc); err != nil {
 				log.Printf("E! Service for input %s failed to start, exiting\n%s\n",
 					input.Name(), err.Error())
@@ -383,21 +398,6 @@ func (a *Agent) Run(shutdown chan struct{}) error {
 	}
 
 	wg.Wait()
+	a.Close()
 	return nil
 }
-
-func copyMetric(m telegraf.Metric) telegraf.Metric {
-	t := time.Time(m.Time())
-
-	tags := make(map[string]string)
-	fields := make(map[string]interface{})
-	for k, v := range m.Tags() {
-		tags[k] = v
-	}
-	for k, v := range m.Fields() {
-		fields[k] = v
-	}
-
-	out, _ := telegraf.NewMetric(m.Name(), tags, fields, t)
-	return out
-}

circle.yml

@@ -1,13 +1,11 @@
 machine:
+  go:
+    version: 1.8.1
   services:
     - docker
-  post:
-    - sudo service zookeeper stop
-    - go version
-    - go version | grep 1.7.3 || sudo rm -rf /usr/local/go
-    - wget https://storage.googleapis.com/golang/go1.7.3.linux-amd64.tar.gz
-    - sudo tar -C /usr/local -xzf go1.7.3.linux-amd64.tar.gz
-    - go version
+    - memcached
+    - redis
+    - rabbitmq-server
 
 dependencies:
   override:

cmd/telegraf/telegraf.go

@@ -4,6 +4,8 @@ import (
 	"flag"
 	"fmt"
 	"log"
+	"net/http"
+	_ "net/http/pprof" // Comment this line to disable pprof endpoint.
 	"os"
 	"os/signal"
 	"runtime"
@@ -24,6 +26,8 @@ import (
 
 var fDebug = flag.Bool("debug", false,
 	"turn on debug logging")
+var pprofAddr = flag.String("pprof-addr", "",
+	"pprof address to listen on, not activate pprof if empty")
 var fQuiet = flag.Bool("quiet", false,
 	"run in quiet mode")
 var fTest = flag.Bool("test", false, "gather metrics, print them out, and exit")
@@ -87,6 +91,7 @@ The commands & flags are:
   --output-filter     filter the output plugins to enable, separator is :
   --usage             print usage for a plugin, ie, 'telegraf --usage mysql'
   --debug             print metrics as they're generated to stdout
+  --pprof-addr        pprof address to listen on, format: localhost:6060 or :6060
   --quiet             run in quiet mode
 
 Examples:
@@ -105,98 +110,24 @@ Examples:
 
   # run telegraf, enabling the cpu & memory input, and influxdb output plugins
   telegraf --config telegraf.conf --input-filter cpu:mem --output-filter influxdb
+
+  # run telegraf with pprof
+  telegraf --config telegraf.conf --pprof-addr localhost:6060
 `
 
 var stop chan struct{}
 
-var srvc service.Service
-
-type program struct{}
-
-func reloadLoop(stop chan struct{}, s service.Service) {
-	defer func() {
-		if service.Interactive() {
-			os.Exit(0)
-		}
-		return
-	}()
+func reloadLoop(
+	stop chan struct{},
+	inputFilters []string,
+	outputFilters []string,
+	aggregatorFilters []string,
+	processorFilters []string,
+) {
 	reload := make(chan bool, 1)
 	reload <- true
 	for <-reload {
 		reload <- false
-		flag.Parse()
-		args := flag.Args()
-
-		var inputFilters []string
-		if *fInputFilters != "" {
-			inputFilter := strings.TrimSpace(*fInputFilters)
-			inputFilters = strings.Split(":"+inputFilter+":", ":")
-		}
-		var outputFilters []string
-		if *fOutputFilters != "" {
-			outputFilter := strings.TrimSpace(*fOutputFilters)
-			outputFilters = strings.Split(":"+outputFilter+":", ":")
-		}
-		var aggregatorFilters []string
-		if *fAggregatorFilters != "" {
-			aggregatorFilter := strings.TrimSpace(*fAggregatorFilters)
-			aggregatorFilters = strings.Split(":"+aggregatorFilter+":", ":")
-		}
-		var processorFilters []string
-		if *fProcessorFilters != "" {
-			processorFilter := strings.TrimSpace(*fProcessorFilters)
-			processorFilters = strings.Split(":"+processorFilter+":", ":")
-		}
-
-		if len(args) > 0 {
-			switch args[0] {
-			case "version":
-				fmt.Printf("Telegraf v%s (git: %s %s)\n", version, branch, commit)
-				return
-			case "config":
-				config.PrintSampleConfig(
-					inputFilters,
-					outputFilters,
-					aggregatorFilters,
-					processorFilters,
-				)
-				return
-			}
-		}
-
-		// switch for flags which just do something and exit immediately
-		switch {
-		case *fOutputList:
-			fmt.Println("Available Output Plugins:")
-			for k, _ := range outputs.Outputs {
-				fmt.Printf("  %s\n", k)
-			}
-			return
-		case *fInputList:
-			fmt.Println("Available Input Plugins:")
-			for k, _ := range inputs.Inputs {
-				fmt.Printf("  %s\n", k)
-			}
-			return
-		case *fVersion:
-			fmt.Printf("Telegraf v%s (git: %s %s)\n", version, branch, commit)
-			return
-		case *fSampleConfig:
-			config.PrintSampleConfig(
-				inputFilters,
-				outputFilters,
-				aggregatorFilters,
-				processorFilters,
-			)
-			return
-		case *fUsage != "":
-			if err := config.PrintInputConfig(*fUsage); err != nil {
-				if err2 := config.PrintOutputConfig(*fUsage); err2 != nil {
-					log.Fatalf("E! %s and %s", err, err2)
-				}
-			}
-			return
-		}
 
 		// If no other options are specified, load the config file and run.
 		c := config.NewConfig()
@@ -213,7 +144,7 @@ func reloadLoop(stop chan struct{}, s service.Service) {
 				log.Fatal("E! " + err.Error())
 			}
 		}
-		if len(c.Outputs) == 0 {
+		if !*fTest && len(c.Outputs) == 0 {
 			log.Fatalf("E! Error: no outputs found, did you provide a valid config file?")
 		}
 		if len(c.Inputs) == 0 {
@@ -237,7 +168,7 @@ func reloadLoop(stop chan struct{}, s service.Service) {
 			if err != nil {
 				log.Fatal("E! " + err.Error())
 			}
-			return
+			os.Exit(0)
 		}
 
 		err = ag.Connect()
@@ -271,14 +202,21 @@ func reloadLoop(stop chan struct{}, s service.Service) {
 		log.Printf("I! Tags enabled: %s", c.ListTags())
 
 		if *fPidfile != "" {
-			f, err := os.Create(*fPidfile)
+			f, err := os.OpenFile(*fPidfile, os.O_CREATE|os.O_WRONLY, 0644)
 			if err != nil {
-				log.Fatalf("E! Unable to create pidfile: %s", err)
+				log.Printf("E! Unable to create pidfile: %s", err)
+			} else {
+				fmt.Fprintf(f, "%d\n", os.Getpid())
+
+				f.Close()
+
+				defer func() {
+					err := os.Remove(*fPidfile)
+					if err != nil {
+						log.Printf("E! Unable to remove pidfile: %s", err)
+					}
+				}()
 			}
-
-			fmt.Fprintf(f, "%d\n", os.Getpid())
-
-			f.Close()
 		}
 
 		ag.Run(shutdown)
@@ -290,14 +228,26 @@ func usageExit(rc int) {
 	os.Exit(rc)
 }
 
+type program struct {
+	inputFilters      []string
+	outputFilters     []string
+	aggregatorFilters []string
+	processorFilters  []string
+}
+
 func (p *program) Start(s service.Service) error {
-	srvc = s
 	go p.run()
 	return nil
 }
 func (p *program) run() {
 	stop = make(chan struct{})
-	reloadLoop(stop, srvc)
+	reloadLoop(
+		stop,
+		p.inputFilters,
+		p.outputFilters,
+		p.aggregatorFilters,
+		p.processorFilters,
+	)
 }
 func (p *program) Stop(s service.Service) error {
 	close(stop)
@@ -307,6 +257,91 @@ func (p *program) Stop(s service.Service) error {
 func main() {
 	flag.Usage = func() { usageExit(0) }
 	flag.Parse()
+	args := flag.Args()
+
+	inputFilters, outputFilters := []string{}, []string{}
+	if *fInputFilters != "" {
+		inputFilters = strings.Split(":"+strings.TrimSpace(*fInputFilters)+":", ":")
+	}
+	if *fOutputFilters != "" {
+		outputFilters = strings.Split(":"+strings.TrimSpace(*fOutputFilters)+":", ":")
+	}
+
+	aggregatorFilters, processorFilters := []string{}, []string{}
+	if *fAggregatorFilters != "" {
+		aggregatorFilters = strings.Split(":"+strings.TrimSpace(*fAggregatorFilters)+":", ":")
+	}
+	if *fProcessorFilters != "" {
+		processorFilters = strings.Split(":"+strings.TrimSpace(*fProcessorFilters)+":", ":")
+	}
+
+	if *pprofAddr != "" {
+		go func() {
+			pprofHostPort := *pprofAddr
+			parts := strings.Split(pprofHostPort, ":")
+			if len(parts) == 2 && parts[0] == "" {
+				pprofHostPort = fmt.Sprintf("localhost:%s", parts[1])
+			}
+			pprofHostPort = "http://" + pprofHostPort + "/debug/pprof"
+
+			log.Printf("I! Starting pprof HTTP server at: %s", pprofHostPort)
+
+			if err := http.ListenAndServe(*pprofAddr, nil); err != nil {
+				log.Fatal("E! " + err.Error())
+			}
+		}()
+	}
+
+	if len(args) > 0 {
+		switch args[0] {
+		case "version":
+			fmt.Printf("Telegraf v%s (git: %s %s)\n", version, branch, commit)
+			return
+		case "config":
+			config.PrintSampleConfig(
+				inputFilters,
+				outputFilters,
+				aggregatorFilters,
+				processorFilters,
+			)
+			return
+		}
+	}
+
+	// switch for flags which just do something and exit immediately
+	switch {
+	case *fOutputList:
+		fmt.Println("Available Output Plugins:")
+		for k, _ := range outputs.Outputs {
+			fmt.Printf("  %s\n", k)
+		}
+		return
+	case *fInputList:
+		fmt.Println("Available Input Plugins:")
+		for k, _ := range inputs.Inputs {
+			fmt.Printf("  %s\n", k)
+		}
+		return
+	case *fVersion:
+		fmt.Printf("Telegraf v%s (git: %s %s)\n", version, branch, commit)
+		return
+	case *fSampleConfig:
+		config.PrintSampleConfig(
+			inputFilters,
+			outputFilters,
+			aggregatorFilters,
+			processorFilters,
+		)
+		return
+	case *fUsage != "":
+		err := config.PrintInputConfig(*fUsage)
+		err2 := config.PrintOutputConfig(*fUsage)
+		if err != nil && err2 != nil {
+			log.Fatalf("E! %s and %s", err, err2)
+		}
+		return
+	}
+
 	if runtime.GOOS == "windows" {
 		svcConfig := &service.Config{
 			Name:        "telegraf",
@@ -316,7 +351,12 @@ func main() {
 			Arguments: []string{"-config", "C:\\Program Files\\Telegraf\\telegraf.conf"},
 		}
 
-		prg := &program{}
+		prg := &program{
+			inputFilters:      inputFilters,
+			outputFilters:     outputFilters,
+			aggregatorFilters: aggregatorFilters,
+			processorFilters:  processorFilters,
+		}
 		s, err := service.New(prg, svcConfig)
 		if err != nil {
 			log.Fatal("E! " + err.Error())
@@ -327,10 +367,14 @@ func main() {
 			if *fConfig != "" {
 				(*svcConfig).Arguments = []string{"-config", *fConfig}
 			}
+			if *fConfigDirectory != "" {
+				(*svcConfig).Arguments = append((*svcConfig).Arguments, "-config-directory", *fConfigDirectory)
+			}
 			err := service.Control(s, *fService)
 			if err != nil {
 				log.Fatal("E! " + err.Error())
 			}
+			os.Exit(0)
 		} else {
 			err = s.Run()
 			if err != nil {
@@ -339,6 +383,12 @@ func main() {
 		}
 	} else {
 		stop = make(chan struct{})
-		reloadLoop(stop, nil)
+		reloadLoop(
+			stop,
+			inputFilters,
+			outputFilters,
+			aggregatorFilters,
+			processorFilters,
+		)
 	}
 }

docs/AGGREGATORS_AND_PROCESSORS.md

@@ -0,0 +1,59 @@
+# Telegraf Aggregator & Processor Plugins
+
+As of release 1.1.0, Telegraf has the concept of Aggregator and Processor Plugins.
+
+These plugins sit in-between Input & Output plugins, aggregating and processing
+metrics as they pass through Telegraf:
+
+```
+┌───────────┐
+│           │
+│    CPU    │───┐
+│           │   │
+└───────────┘   │
+                │
+┌───────────┐   │                                              ┌───────────┐
+│           │   │                                              │           │
+│  Memory   │───┤                                          ┌──▶│ InfluxDB  │
+│           │   │                                          │   │           │
+└───────────┘   │    ┌─────────────┐     ┌─────────────┐   │   └───────────┘
+                │    │             │     │Aggregate    │   │
+┌───────────┐   │    │Process      │     │ - mean      │   │   ┌───────────┐
+│           │   │    │ - transform │     │ - quantiles │   │   │           │
+│   MySQL   │───┼───▶│ - decorate  │────▶│ - min/max   │───┼──▶│   File    │
+│           │   │    │ - filter    │     │ - count     │   │   │           │
+└───────────┘   │    │             │     │             │   │   └───────────┘
+                │    └─────────────┘     └─────────────┘   │
+┌───────────┐   │                                          │   ┌───────────┐
+│           │   │                                          │   │           │
+│   SNMP    │───┤                                          └──▶│   Kafka   │
+│           │   │                                              │           │
+└───────────┘   │                                              └───────────┘
+                │
+┌───────────┐   │
+│           │   │
+│  Docker   │───┘
+│           │
+└───────────┘
+```
+
+Both Aggregators and Processors analyze metrics as they pass through Telegraf.
+
+**Processor** plugins process metrics as they pass through and immediately emit
+results based on the values they process. For example, this could be printing
+all metrics or adding a tag to all metrics that pass through.
+
+**Aggregator** plugins, on the other hand, are a bit more complicated. Aggregators
+are typically for emitting new _aggregate_ metrics, such as a running mean,
+minimum, maximum, quantiles, or standard deviation. For this reason, all _aggregator_
+plugins are configured with a `period`. The `period` is the size of the window
+of metrics that each _aggregate_ represents. In other words, the emitted
+_aggregate_ metric will be the aggregated value of the past `period` seconds.
+Since many users will only care about their aggregates and not every single metric
+gathered, there is also a `drop_original` argument, which tells Telegraf to only
+emit the aggregates and not the original metrics.
+
+**NOTE** That since aggregators only aggregate metrics within their period, that
+historical data is not supported. In other words, if your metric timestamp is more
+than `now() - period` in the past, it will not be aggregated. If this is a feature
+that you need, please comment on this [github issue](https://github.com/influxdata/telegraf/issues/1992)

docs/CONFIGURATION.md

@@ -24,6 +24,16 @@ Environment variables can be used anywhere in the config file, simply prepend
 them with $. For strings the variable must be within quotes (ie, "$STR_VAR"),
 for numbers and booleans they should be plain (ie, $INT_VAR, $BOOL_VAR)
 
+## Configuration file locations
+
+The location of the configuration file can be set via the `--config` command
+line flag. Telegraf will also pick up all files matching the pattern `*.conf` if
+the `-config-directory` command line flag is used.
+
+On most systems, the default locations are `/etc/telegraf/telegraf.conf` for
+the main configuration file and `/etc/telegraf/telegraf.d` for the directory of
+configuration files.
+
 # Global Tags
 
 Global tags can be specified in the `[global_tags]` section of the config file
@@ -60,7 +70,7 @@ ie, a jitter of 5s and flush_interval 10s means flushes will happen every 10-15s
 as the collection interval, with the maximum being 1s. Precision will NOT
 be used for service inputs, such as logparser and statsd. Valid values are
 "ns", "us" (or "µs"), "ms", "s".
-* **logfile**: Specify the log file name. The empty string means to log to stdout.
+* **logfile**: Specify the log file name. The empty string means to log to stderr.
 * **debug**: Run telegraf in debug mode.
 * **quiet**: Run telegraf in quiet mode (error messages only).
 * **hostname**: Override default hostname, if empty use os.Hostname().
@@ -114,31 +124,40 @@ is not specified then processor execution order will be random.
 Filters can be configured per input, output, processor, or aggregator,
 see below for examples.
 
-* **namepass**: An array of strings that is used to filter metrics generated by the
-current input. Each string in the array is tested as a glob match against
-measurement names and if it matches, the field is emitted.
-* **namedrop**: The inverse of pass, if a measurement name matches, it is not emitted.
-* **fieldpass**: An array of strings that is used to filter metrics generated by the
-current input. Each string in the array is tested as a glob match against field names
-and if it matches, the field is emitted. fieldpass is not available for outputs.
-* **fielddrop**: The inverse of pass, if a field name matches, it is not emitted.
-fielddrop is not available for outputs.
-* **tagpass**: tag names and arrays of strings that are used to filter
-measurements by the current input. Each string in the array is tested as a glob
-match against the tag name, and if it matches the measurement is emitted.
-* **tagdrop**: The inverse of tagpass. If a tag matches, the measurement is not
-emitted. This is tested on measurements that have passed the tagpass test.
-* **tagexclude**: tagexclude can be used to exclude a tag from measurement(s).
-As opposed to tagdrop, which will drop an entire measurement based on it's
-tags, tagexclude simply strips the given tag keys from the measurement. This
-can be used on inputs & outputs, but it is _recommended_ to be used on inputs,
-as it is more efficient to filter out tags at the ingestion point.
-* **taginclude**: taginclude is the inverse of tagexclude. It will only include
-the tag keys in the final measurement.
+* **namepass**:
+An array of glob pattern strings.  Only points whose measurement name matches
+a pattern in this list are emitted.
+* **namedrop**:
+The inverse of `namepass`.  If a match is found the point is discarded. This
+is tested on points after they have passed the `namepass` test.
+* **fieldpass**:
+An array of glob pattern strings.  Only fields whose field key matches a
+pattern in this list are emitted.  Not available for outputs.
+* **fielddrop**:
+The inverse of `fieldpass`. Fields with a field key matching one of the
+patterns will be discarded from the point.  Not available for outputs.
+* **tagpass**:
+A table mapping tag keys to arrays of glob pattern strings.  Only points
+that contain a tag key in the table and a tag value matching one of its
+patterns is emitted.
+* **tagdrop**:
+The inverse of `tagpass`.  If a match is found the point is discarded. This
+is tested on points after they have passed the `tagpass` test.
+* **taginclude**:
+An array of glob pattern strings.  Only tags with a tag key matching one of
+the patterns are emitted.  In contrast to `tagpass`, which will pass an entire
+point based on its tag, `taginclude` removes all non matching tags from the
+point.  This filter can be used on both inputs & outputs, but it is
+_recommended_ to be used on inputs, as it is more efficient to filter out tags
+at the ingestion point.
+* **tagexclude**:
+The inverse of `taginclude`. Tags with a tag key matching one of the patterns
+will be discarded from the point.
 
-**NOTE** `tagpass` and `tagdrop` parameters must be defined at the _end_ of
-the plugin definition, otherwise subsequent plugin config options will be
-interpreted as part of the tagpass/tagdrop map.
+**NOTE** Due to the way TOML is parsed, `tagpass` and `tagdrop` parameters
+must be defined at the _end_ of the plugin definition, otherwise subsequent
+plugin config options will be interpreted as part of the tagpass/tagdrop
+tables.
 
 #### Input Configuration Examples
 
@@ -351,4 +370,4 @@ to the system load metrics due to the `namepass` parameter.
 
 [[outputs.file]]
   files = ["stdout"]
-```
+```

docs/DATA_FORMATS_INPUT.md

@@ -7,6 +7,7 @@ Telegraf is able to parse the following input data formats into metrics:
 1. [Graphite](https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md#graphite)
 1. [Value](https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md#value), ie: 45 or "booyah"
 1. [Nagios](https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md#nagios) (exec input only)
+1. [Collectd](https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md#collectd)
 
 Telegraf metrics, like InfluxDB
 [points](https://docs.influxdata.com/influxdb/v0.10/write_protocols/line/),
@@ -40,7 +41,7 @@ example, in the exec plugin:
   name_suffix = "_mycollector"
 
   ## Data format to consume.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
   data_format = "json"
@@ -67,7 +68,7 @@ metrics are parsed directly into Telegraf metrics.
   name_suffix = "_mycollector"
 
   ## Data format to consume.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
   data_format = "influx"
@@ -117,7 +118,7 @@ For example, if you had this configuration:
   name_suffix = "_mycollector"
 
   ## Data format to consume.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
   data_format = "json"
@@ -147,6 +148,62 @@ Your Telegraf metrics would get tagged with "my_tag_1"
 exec_mycollector,my_tag_1=foo a=5,b_c=6
 ```
 
+If the JSON data is an array, then each element of the array is parsed with the configured settings.
+Each resulting metric will be output with the same timestamp.
+
+For example, if the following configuration:
+
+```toml
+[[inputs.exec]]
+  ## Commands array
+  commands = ["/usr/bin/mycollector --foo=bar"]
+
+  ## measurement name suffix (for separating different commands)
+  name_suffix = "_mycollector"
+
+  ## Data format to consume.
+  ## Each data format has its own unique set of configuration options, read
+  ## more about them here:
+  ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
+  data_format = "json"
+
+  ## List of tag names to extract from top-level of JSON server response
+  tag_keys = [
+    "my_tag_1",
+    "my_tag_2"
+  ]
+```
+
+with this JSON output from a command:
+
+```json
+[
+    {
+        "a": 5,
+        "b": {
+            "c": 6
+        },
+        "my_tag_1": "foo",
+        "my_tag_2": "baz"
+    },
+    {
+        "a": 7,
+        "b": {
+            "c": 8
+        },
+        "my_tag_1": "bar",
+        "my_tag_2": "baz"
+    }
+]
+```
+
+Your Telegraf metrics would get tagged with "my_tag_1" and "my_tag_2"
+
+```
+exec_mycollector,my_tag_1=foo,my_tag_2=baz a=5,b_c=6
+exec_mycollector,my_tag_1=bar,my_tag_2=baz a=7,b_c=8
+```
+
 # Value:
 
 The "value" data format translates single values into Telegraf metrics. This
@@ -176,7 +233,7 @@ name of the plugin.
   name_override = "entropy_available"
 
   ## Data format to consume.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
   data_format = "value"
@@ -334,7 +391,7 @@ There are many more options available,
   name_suffix = "_mycollector"
 
   ## Data format to consume.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
   data_format = "graphite"
@@ -371,14 +428,54 @@ Note: Nagios Input Data Formats is only supported in `exec` input plugin.
 ```toml
 [[inputs.exec]]
   ## Commands array
-  commands = ["/usr/lib/nagios/plugins/check_load", "-w 5,6,7 -c 7,8,9"]
+  commands = ["/usr/lib/nagios/plugins/check_load -w 5,6,7 -c 7,8,9"]
 
   ## measurement name suffix (for separating different commands)
   name_suffix = "_mycollector"
 
   ## Data format to consume.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
   data_format = "nagios"
 ```
+
+# Collectd:
+
+The collectd format parses the collectd binary network protocol.  Tags are
+created for host, instance, type, and type instance.  All collectd values are
+added as float64 fields.
+
+For more information about the binary network protocol see
+[here](https://collectd.org/wiki/index.php/Binary_protocol).
+
+You can control the cryptographic settings with parser options.  Create an
+authentication file and set `collectd_auth_file` to the path of the file, then
+set the desired security level in `collectd_security_level`.
+
+Additional information including client setup can be found
+[here](https://collectd.org/wiki/index.php/Networking_introduction#Cryptographic_setup).
+
+You can also change the path to the typesdb or add additional typesdb using
+`collectd_typesdb`.
+
+#### Collectd Configuration:
+
+```toml
+[[inputs.socket_listener]]
+  service_address = "udp://127.0.0.1:25826"
+  name_prefix = "collectd_"
+
+  ## Data format to consume.
+  ## Each data format has its own unique set of configuration options, read
+  ## more about them here:
+  ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
+  data_format = "collectd"
+
+  ## Authentication file for cryptographic security levels
+  collectd_auth_file = "/etc/collectd/auth_file"
+  ## One of none (default), sign, or encrypt
+  collectd_security_level = "encrypt"
+  ## Path of to TypesDB specifications
+  collectd_typesdb = ["/usr/share/collectd/types.db"]
+```

docs/DATA_FORMATS_OUTPUT.md

@@ -36,7 +36,7 @@ config option, for example, in the `file` output plugin:
   files = ["stdout"]
 
   ## Data format to output.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_OUTPUT.md
   data_format = "influx"
@@ -60,7 +60,7 @@ metrics are serialized directly into InfluxDB line-protocol.
   files = ["stdout", "/tmp/metrics.out"]
 
   ## Data format to output.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_OUTPUT.md
   data_format = "influx"
@@ -104,7 +104,7 @@ tars.cpu-total.us-east-1.cpu.usage_idle 98.09 1455320690
   files = ["stdout", "/tmp/metrics.out"]
 
   ## Data format to output.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_OUTPUT.md
   data_format = "graphite"
@@ -143,8 +143,18 @@ The JSON data format serialized Telegraf metrics in json format. The format is:
   files = ["stdout", "/tmp/metrics.out"]
 
   ## Data format to output.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_OUTPUT.md
   data_format = "json"
+  json_timestamp_units = "1ns"
 ```
+
+By default, the timestamp that is output in JSON data format serialized Telegraf
+metrics is in seconds. The precision of this timestamp can be adjusted for any output
+by adding the optional `json_timestamp_units` parameter to the configuration for
+that output. This parameter can be used to set the timestamp units to  nanoseconds (`ns`),
+microseconds (`us` or `µs`), milliseconds (`ms`), or seconds (`s`). Note that this
+parameter will be truncated to the nearest power of 10 that, so if the `json_timestamp_units`
+are set to `15ms` the timestamps for the JSON format serialized Telegraf metrics will be
+output in hundredths of a second (`10ms`).

docs/LICENSE_OF_DEPENDENCIES.md

@@ -1,4 +1,5 @@
 # List
+- collectd.org [MIT LICENSE](https://github.com/collectd/go-collectd/blob/master/LICENSE)
 - github.com/Shopify/sarama [MIT LICENSE](https://github.com/Shopify/sarama/blob/master/MIT-LICENSE)
 - github.com/Sirupsen/logrus [MIT LICENSE](https://github.com/Sirupsen/logrus/blob/master/LICENSE)
 - github.com/armon/go-metrics [MIT LICENSE](https://github.com/armon/go-metrics/blob/master/LICENSE)
@@ -12,11 +13,11 @@
 - github.com/gogo/protobuf [BSD LICENSE](https://github.com/gogo/protobuf/blob/master/LICENSE)
 - github.com/golang/protobuf [BSD LICENSE](https://github.com/golang/protobuf/blob/master/LICENSE)
 - github.com/golang/snappy [BSD LICENSE](https://github.com/golang/snappy/blob/master/LICENSE)
-- github.com/gonuts/go-shellquote (No License, but the project it was forked from https://github.com/kballard/go-shellquote is [MIT](https://github.com/kballard/go-shellquote/blob/master/LICENSE)).
 - github.com/hashicorp/go-msgpack [BSD LICENSE](https://github.com/hashicorp/go-msgpack/blob/master/LICENSE)
 - github.com/hashicorp/raft [MPL LICENSE](https://github.com/hashicorp/raft/blob/master/LICENSE)
 - github.com/hashicorp/raft-boltdb [MPL LICENSE](https://github.com/hashicorp/raft-boltdb/blob/master/LICENSE)
 - github.com/kardianos/service [ZLIB LICENSE](https://github.com/kardianos/service/blob/master/LICENSE) (License not named but matches word for word with ZLib)
+- github.com/kballard/go-shellquote [MIT LICENSE](https://github.com/kballard/go-shellquote/blob/master/LICENSE)
 - github.com/lib/pq [MIT LICENSE](https://github.com/lib/pq/blob/master/LICENSE.md)
 - github.com/matttproud/golang_protobuf_extensions [APACHE LICENSE](https://github.com/matttproud/golang_protobuf_extensions/blob/master/LICENSE)
 - github.com/naoina/go-stringutil [MIT LICENSE](https://github.com/naoina/go-stringutil/blob/master/LICENSE)
@@ -30,4 +31,3 @@
 - gopkg.in/dancannon/gorethink.v1 [APACHE LICENSE](https://github.com/dancannon/gorethink/blob/v1.1.2/LICENSE)
 - gopkg.in/mgo.v2 [BSD LICENSE](https://github.com/go-mgo/mgo/blob/v2/LICENSE)
 - golang.org/x/crypto/ [BSD LICENSE](https://github.com/golang/crypto/blob/master/LICENSE)
-

docs/PROFILING.md

@@ -0,0 +1,24 @@
+# Telegraf profiling
+
+Telegraf uses the standard package `net/http/pprof`. This package serves via its HTTP server runtime profiling data in the format expected by the pprof visualization tool.
+
+By default, the profiling is turned off.
+
+To enable profiling you need to specify address to config parameter `pprof-addr`, for example:
+
+```
+telegraf --config telegraf.conf --pprof-addr localhost:6060
+```
+
+There are several paths to get different profiling information:
+
+To look at the heap profile:
+
+`go tool pprof http://localhost:6060/debug/pprof/heap`
+
+or to look at a 30-second CPU profile:
+
+`go tool pprof http://localhost:6060/debug/pprof/profile?seconds=30`
+
+To view all available profiles, open `http://localhost:6060/debug/pprof/` in your browser.
+

etc/telegraf_windows.conf

@@ -105,10 +105,11 @@
       "% Privileged Time",
       "% User Time",
       "% Processor Time",
+      "% DPC Time",
     ]
     Measurement = "win_cpu"
     # Set to true to include _Total instance when querying for all (*).
-    #IncludeTotal=false
+    IncludeTotal=true
 
   [[inputs.win_perf_counters.object]]
     # Disk times and queues
@@ -116,20 +117,54 @@
     Instances = ["*"]
     Counters = [
       "% Idle Time",
-      "% Disk Time","% Disk Read Time",
+      "% Disk Time",
+      "% Disk Read Time",
       "% Disk Write Time",
-      "% User Time",
       "Current Disk Queue Length",
+      "% Free Space",
+      "Free Megabytes",
     ]
     Measurement = "win_disk"
     # Set to true to include _Total instance when querying for all (*).
     #IncludeTotal=false
 
+  [[inputs.win_perf_counters.object]]
+    ObjectName = "PhysicalDisk"
+    Instances = ["*"]
+    Counters = [
+      "Disk Read Bytes/sec",
+      "Disk Write Bytes/sec",
+      "Current Disk Queue Length",
+      "Disk Reads/sec",
+      "Disk Writes/sec",
+      "% Disk Time",
+      "% Disk Read Time",
+      "% Disk Write Time",
+    ]
+    Measurement = "win_diskio"
+
+  [[inputs.win_perf_counters.object]]
+    ObjectName = "Network Interface"
+    Instances = ["*"]
+    Counters = [
+      "Bytes Received/sec",
+      "Bytes Sent/sec",
+      "Packets Received/sec",
+      "Packets Sent/sec",
+      "Packets Received Discarded",
+      "Packets Outbound Discarded",
+      "Packets Received Errors",
+      "Packets Outbound Errors",
+    ]
+    Measurement = "win_net"
+
   [[inputs.win_perf_counters.object]]
     ObjectName = "System"
     Counters = [
       "Context Switches/sec",
       "System Calls/sec",
+      "Processor Queue Length",
+      "System Up Time",
     ]
     Instances = ["------"]
     Measurement = "win_system"
@@ -149,6 +184,10 @@
       "Transition Faults/sec",
       "Pool Nonpaged Bytes",
       "Pool Paged Bytes",
+      "Standby Cache Reserve Bytes",
+      "Standby Cache Normal Priority Bytes",
+      "Standby Cache Core Bytes",
+
     ]
     # Use 6 x - to remove the Instance bit from the query.
     Instances = ["------"]
@@ -156,6 +195,31 @@
     # Set to true to include _Total instance when querying for all (*).
     #IncludeTotal=false
 
+  [[inputs.win_perf_counters.object]]
+    # Example query where the Instance portion must be removed to get data back,
+    # such as from the Paging File object.
+    ObjectName = "Paging File"
+    Counters = [
+      "% Usage",
+    ]
+    Instances = ["_Total"]
+    Measurement = "win_swap"
+
+  [[inputs.win_perf_counters.object]]
+    ObjectName = "Network Interface"
+    Instances = ["*"]
+    Counters = [
+      "Bytes Sent/sec",
+      "Bytes Received/sec",
+      "Packets Sent/sec",
+      "Packets Received/sec",
+      "Packets Received Discarded",
+      "Packets Received Errors",
+      "Packets Outbound Discarded",
+      "Packets Outbound Errors",
+    ]
+
+
 
 # Windows system plugins using WMI (disabled by default, using
 # win_perf_counters over WMI is recommended)

internal/buffer/buffer.go

@@ -4,15 +4,17 @@ import (
 	"sync"
 
 	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/selfstat"
+)
+
+var (
+	MetricsWritten = selfstat.Register("agent", "metrics_written", map[string]string{})
+	MetricsDropped = selfstat.Register("agent", "metrics_dropped", map[string]string{})
 )
 
 // Buffer is an object for storing metrics in a circular buffer.
 type Buffer struct {
 	buf chan telegraf.Metric
-	// total dropped metrics
-	drops int
-	// total metrics added
-	total int
 
 	mu sync.Mutex
 }
@@ -36,27 +38,18 @@ func (b *Buffer) Len() int {
 	return len(b.buf)
 }
 
-// Drops returns the total number of dropped metrics that have occured in this
-// buffer since instantiation.
-func (b *Buffer) Drops() int {
-	return b.drops
-}
-
-// Total returns the total number of metrics that have been added to this buffer.
-func (b *Buffer) Total() int {
-	return b.total
-}
-
 // Add adds metrics to the buffer.
 func (b *Buffer) Add(metrics ...telegraf.Metric) {
 	for i, _ := range metrics {
-		b.total++
+		MetricsWritten.Incr(1)
 		select {
 		case b.buf <- metrics[i]:
 		default:
-			b.drops++
+			b.mu.Lock()
+			MetricsDropped.Incr(1)
 			<-b.buf
 			b.buf <- metrics[i]
+			b.mu.Unlock()
 		}
 	}
 }

internal/buffer/buffer_test.go

@@ -27,47 +27,53 @@ func BenchmarkAddMetrics(b *testing.B) {
 
 func TestNewBufferBasicFuncs(t *testing.T) {
 	b := NewBuffer(10)
+	MetricsDropped.Set(0)
+	MetricsWritten.Set(0)
 
 	assert.True(t, b.IsEmpty())
 	assert.Zero(t, b.Len())
-	assert.Zero(t, b.Drops())
-	assert.Zero(t, b.Total())
+	assert.Zero(t, MetricsDropped.Get())
+	assert.Zero(t, MetricsWritten.Get())
 
 	m := testutil.TestMetric(1, "mymetric")
 	b.Add(m)
 	assert.False(t, b.IsEmpty())
 	assert.Equal(t, b.Len(), 1)
-	assert.Equal(t, b.Drops(), 0)
-	assert.Equal(t, b.Total(), 1)
+	assert.Equal(t, int64(0), MetricsDropped.Get())
+	assert.Equal(t, int64(1), MetricsWritten.Get())
 
 	b.Add(metricList...)
 	assert.False(t, b.IsEmpty())
 	assert.Equal(t, b.Len(), 6)
-	assert.Equal(t, b.Drops(), 0)
-	assert.Equal(t, b.Total(), 6)
+	assert.Equal(t, int64(0), MetricsDropped.Get())
+	assert.Equal(t, int64(6), MetricsWritten.Get())
 }
 
 func TestDroppingMetrics(t *testing.T) {
 	b := NewBuffer(10)
+	MetricsDropped.Set(0)
+	MetricsWritten.Set(0)
 
 	// Add up to the size of the buffer
 	b.Add(metricList...)
 	b.Add(metricList...)
 	assert.False(t, b.IsEmpty())
 	assert.Equal(t, b.Len(), 10)
-	assert.Equal(t, b.Drops(), 0)
-	assert.Equal(t, b.Total(), 10)
+	assert.Equal(t, int64(0), MetricsDropped.Get())
+	assert.Equal(t, int64(10), MetricsWritten.Get())
 
 	// Add 5 more and verify they were dropped
 	b.Add(metricList...)
 	assert.False(t, b.IsEmpty())
 	assert.Equal(t, b.Len(), 10)
-	assert.Equal(t, b.Drops(), 5)
-	assert.Equal(t, b.Total(), 15)
+	assert.Equal(t, int64(5), MetricsDropped.Get())
+	assert.Equal(t, int64(15), MetricsWritten.Get())
 }
 
 func TestGettingBatches(t *testing.T) {
 	b := NewBuffer(20)
+	MetricsDropped.Set(0)
+	MetricsWritten.Set(0)
 
 	// Verify that the buffer returned is smaller than requested when there are
 	// not as many items as requested.
@@ -78,8 +84,8 @@ func TestGettingBatches(t *testing.T) {
 	// Verify that the buffer is now empty
 	assert.True(t, b.IsEmpty())
 	assert.Zero(t, b.Len())
-	assert.Zero(t, b.Drops())
-	assert.Equal(t, b.Total(), 5)
+	assert.Zero(t, MetricsDropped.Get())
+	assert.Equal(t, int64(5), MetricsWritten.Get())
 
 	// Verify that the buffer returned is not more than the size requested
 	b.Add(metricList...)
@@ -89,6 +95,6 @@ func TestGettingBatches(t *testing.T) {
 	// Verify that buffer is not empty
 	assert.False(t, b.IsEmpty())
 	assert.Equal(t, b.Len(), 2)
-	assert.Equal(t, b.Drops(), 0)
-	assert.Equal(t, b.Total(), 10)
+	assert.Equal(t, int64(0), MetricsDropped.Get())
+	assert.Equal(t, int64(10), MetricsWritten.Get())
 }

internal/config/config.go

@@ -6,6 +6,7 @@ import (
 	"fmt"
 	"io/ioutil"
 	"log"
+	"math"
 	"os"
 	"path/filepath"
 	"regexp"
@@ -25,7 +26,6 @@ import (
 	"github.com/influxdata/telegraf/plugins/processors"
 	"github.com/influxdata/telegraf/plugins/serializers"
 
-	"github.com/influxdata/config"
 	"github.com/influxdata/toml"
 	"github.com/influxdata/toml/ast"
 )
@@ -85,8 +85,8 @@ type AgentConfig struct {
 	//     ie, if Interval=10s then always collect on :00, :10, :20, etc.
 	RoundInterval bool
 
-	// By default, precision will be set to the same timestamp order as the
-	// collection interval, with the maximum being 1s.
+	// By default or when set to "0s", precision will be set to the same
+	// timestamp order as the collection interval, with the maximum being 1s.
 	//   ie, when interval = "10s", precision will be "1s"
 	//       when interval = "250ms", precision will be "1ms"
 	// Precision will NOT be used for service inputs. It is up to each individual
@@ -230,10 +230,13 @@ var header = `# Telegraf Configuration
   ## ie, a jitter of 5s and interval 10s means flushes will happen every 10-15s
   flush_jitter = "0s"
 
-  ## By default, precision will be set to the same timestamp order as the
-  ## collection interval, with the maximum being 1s.
-  ## Precision will NOT be used for service inputs, such as logparser and statsd.
-  ## Valid values are "ns", "us" (or "µs"), "ms", "s".
+  ## By default or when set to "0s", precision will be set to the same
+  ## timestamp order as the collection interval, with the maximum being 1s.
+  ##   ie, when interval = "10s", precision will be "1s"
+  ##       when interval = "250ms", precision will be "1ms"
+  ## Precision will NOT be used for service inputs. It is up to each individual
+  ## service input to set the timestamp at the appropriate precision.
+  ## Valid time units are "ns", "us" (or "µs"), "ms", "s".
   precision = ""
 
   ## Logging configuration:
@@ -506,6 +509,10 @@ func PrintOutputConfig(name string) error {
 
 func (c *Config) LoadDirectory(path string) error {
 	walkfn := func(thispath string, info os.FileInfo, _ error) error {
+		if info == nil {
+			log.Printf("W! Telegraf is not permitted to read %s", thispath)
+			return nil
+		}
 		if info.IsDir() {
 			return nil
 		}
@@ -566,7 +573,7 @@ func (c *Config) LoadConfig(path string) error {
 			if !ok {
 				return fmt.Errorf("%s: invalid configuration", path)
 			}
-			if err = config.UnmarshalTable(subTable, c.Tags); err != nil {
+			if err = toml.UnmarshalTable(subTable, c.Tags); err != nil {
 				log.Printf("E! Could not parse [global_tags] config\n")
 				return fmt.Errorf("Error parsing %s, %s", path, err)
 			}
@@ -579,7 +586,7 @@ func (c *Config) LoadConfig(path string) error {
 		if !ok {
 			return fmt.Errorf("%s: invalid configuration", path)
 		}
-		if err = config.UnmarshalTable(subTable, c.Agent); err != nil {
+		if err = toml.UnmarshalTable(subTable, c.Agent); err != nil {
 			log.Printf("E! Could not parse [agent] config\n")
 			return fmt.Errorf("Error parsing %s, %s", path, err)
 		}
@@ -716,7 +723,7 @@ func (c *Config) addAggregator(name string, table *ast.Table) error {
 		return err
 	}
 
-	if err := config.UnmarshalTable(table, aggregator); err != nil {
+	if err := toml.UnmarshalTable(table, aggregator); err != nil {
 		return err
 	}
 
@@ -736,7 +743,7 @@ func (c *Config) addProcessor(name string, table *ast.Table) error {
 		return err
 	}
 
-	if err := config.UnmarshalTable(table, processor); err != nil {
+	if err := toml.UnmarshalTable(table, processor); err != nil {
 		return err
 	}
 
@@ -776,7 +783,7 @@ func (c *Config) addOutput(name string, table *ast.Table) error {
 		return err
 	}
 
-	if err := config.UnmarshalTable(table, output); err != nil {
+	if err := toml.UnmarshalTable(table, output); err != nil {
 		return err
 	}
 
@@ -817,14 +824,11 @@ func (c *Config) addInput(name string, table *ast.Table) error {
 		return err
 	}
 
-	if err := config.UnmarshalTable(table, input); err != nil {
+	if err := toml.UnmarshalTable(table, input); err != nil {
 		return err
 	}
 
-	rp := &models.RunningInput{
-		Input:  input,
-		Config: pluginConfig,
-	}
+	rp := models.NewRunningInput(input, pluginConfig)
 	c.Inputs = append(c.Inputs, rp)
 	return nil
 }
@@ -912,7 +916,7 @@ func buildAggregator(name string, tbl *ast.Table) (*models.AggregatorConfig, err
 	conf.Tags = make(map[string]string)
 	if node, ok := tbl.Fields["tags"]; ok {
 		if subtbl, ok := node.(*ast.Table); ok {
-			if err := config.UnmarshalTable(subtbl, conf.Tags); err != nil {
+			if err := toml.UnmarshalTable(subtbl, conf.Tags); err != nil {
 				log.Printf("Could not parse tags for input %s\n", name)
 			}
 		}
@@ -1149,7 +1153,7 @@ func buildInput(name string, tbl *ast.Table) (*models.InputConfig, error) {
 	cp.Tags = make(map[string]string)
 	if node, ok := tbl.Fields["tags"]; ok {
 		if subtbl, ok := node.(*ast.Table); ok {
-			if err := config.UnmarshalTable(subtbl, cp.Tags); err != nil {
+			if err := toml.UnmarshalTable(subtbl, cp.Tags); err != nil {
 				log.Printf("E! Could not parse tags for input %s\n", name)
 			}
 		}
@@ -1229,6 +1233,34 @@ func buildParser(name string, tbl *ast.Table) (parsers.Parser, error) {
 		}
 	}
 
+	if node, ok := tbl.Fields["collectd_auth_file"]; ok {
+		if kv, ok := node.(*ast.KeyValue); ok {
+			if str, ok := kv.Value.(*ast.String); ok {
+				c.CollectdAuthFile = str.Value
+			}
+		}
+	}
+
+	if node, ok := tbl.Fields["collectd_security_level"]; ok {
+		if kv, ok := node.(*ast.KeyValue); ok {
+			if str, ok := kv.Value.(*ast.String); ok {
+				c.CollectdSecurityLevel = str.Value
+			}
+		}
+	}
+
+	if node, ok := tbl.Fields["collectd_typesdb"]; ok {
+		if kv, ok := node.(*ast.KeyValue); ok {
+			if ary, ok := kv.Value.(*ast.Array); ok {
+				for _, elem := range ary.Value {
+					if str, ok := elem.(*ast.String); ok {
+						c.CollectdTypesDB = append(c.CollectdTypesDB, str.Value)
+					}
+				}
+			}
+		}
+	}
+
 	c.MetricName = name
 
 	delete(tbl.Fields, "data_format")
@@ -1236,6 +1268,9 @@ func buildParser(name string, tbl *ast.Table) (parsers.Parser, error) {
 	delete(tbl.Fields, "templates")
 	delete(tbl.Fields, "tag_keys")
 	delete(tbl.Fields, "data_type")
+	delete(tbl.Fields, "collectd_auth_file")
+	delete(tbl.Fields, "collectd_security_level")
+	delete(tbl.Fields, "collectd_typesdb")
 
 	return parsers.NewParser(c)
 }
@@ -1244,7 +1279,7 @@ func buildParser(name string, tbl *ast.Table) (parsers.Parser, error) {
 // a serializers.Serializer object, and creates it, which can then be added onto
 // an Output object.
 func buildSerializer(name string, tbl *ast.Table) (serializers.Serializer, error) {
-	c := &serializers.Config{}
+	c := &serializers.Config{TimestampUnits: time.Duration(1 * time.Second)}
 
 	if node, ok := tbl.Fields["data_format"]; ok {
 		if kv, ok := node.(*ast.KeyValue); ok {
@@ -1274,9 +1309,26 @@ func buildSerializer(name string, tbl *ast.Table) (serializers.Serializer, error
 		}
 	}
 
+	if node, ok := tbl.Fields["json_timestamp_units"]; ok {
+		if kv, ok := node.(*ast.KeyValue); ok {
+			if str, ok := kv.Value.(*ast.String); ok {
+				timestampVal, err := time.ParseDuration(str.Value)
+				if err != nil {
+					return nil, fmt.Errorf("Unable to parse json_timestamp_units as a duration, %s", err)
+				}
+				// now that we have a duration, truncate it to the nearest
+				// power of ten (just in case)
+				nearest_exponent := int64(math.Log10(float64(timestampVal.Nanoseconds())))
+				new_nanoseconds := int64(math.Pow(10.0, float64(nearest_exponent)))
+				c.TimestampUnits = time.Duration(new_nanoseconds)
+			}
+		}
+	}
+
 	delete(tbl.Fields, "data_format")
 	delete(tbl.Fields, "prefix")
 	delete(tbl.Fields, "template")
+	delete(tbl.Fields, "json_timestamp_units")
 	return serializers.NewSerializer(c)
 }
 

internal/config/testdata/telegraf-agent.toml

@@ -60,7 +60,7 @@
   # Kafka topic for producer messages
   topic = "telegraf"
   # Telegraf tag to use as a routing key
-  #  ie, if this tag exists, it's value will be used as the routing key
+  #  ie, if this tag exists, its value will be used as the routing key
   routing_tag = "host"
 
 

internal/errchan/errchan.go

@@ -1,37 +0,0 @@
-package errchan
-
-import (
-	"fmt"
-	"strings"
-)
-
-type ErrChan struct {
-	C chan error
-}
-
-// New returns an error channel of max length 'n'
-// errors can be sent to the ErrChan.C channel, and will be returned when
-// ErrChan.Error() is called.
-func New(n int) *ErrChan {
-	return &ErrChan{
-		C: make(chan error, n),
-	}
-}
-
-// Error closes the ErrChan.C channel and returns an error if there are any
-// non-nil errors, otherwise returns nil.
-func (e *ErrChan) Error() error {
-	close(e.C)
-
-	var out string
-	for err := range e.C {
-		if err != nil {
-			out += "[" + err.Error() + "], "
-		}
-	}
-
-	if out != "" {
-		return fmt.Errorf("Errors encountered: " + strings.TrimRight(out, ", "))
-	}
-	return nil
-}

internal/globpath/globpath_test.go

@@ -28,7 +28,7 @@ func TestCompileAndMatch(t *testing.T) {
 	require.NoError(t, err)
 
 	matches := g1.Match()
-	assert.Len(t, matches, 3)
+	assert.Len(t, matches, 6)
 	matches = g2.Match()
 	assert.Len(t, matches, 2)
 	matches = g3.Match()
@@ -56,6 +56,16 @@ func TestFindRootDir(t *testing.T) {
 	}
 }
 
+func TestFindNestedTextFile(t *testing.T) {
+	dir := getTestdataDir()
+	// test super asterisk
+	g1, err := Compile(dir + "/**.txt")
+	require.NoError(t, err)
+
+	matches := g1.Match()
+	assert.Len(t, matches, 1)
+}
+
 func getTestdataDir() string {
 	_, filename, _, _ := runtime.Caller(1)
 	return strings.Replace(filename, "globpath_test.go", "testdata", 1)

internal/models/makemetric.go

@@ -6,6 +6,7 @@ import (
 	"time"
 
 	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/metric"
 )
 
 // makemetric is used by both RunningAggregator & RunningInput
@@ -31,7 +32,6 @@ func makemetric(
 	daemonTags map[string]string,
 	filter Filter,
 	applyFilter bool,
-	debug bool,
 	mType telegraf.ValueType,
 	t time.Time,
 ) telegraf.Metric {
@@ -122,11 +122,9 @@ func makemetric(
 		case float64:
 			// NaNs are invalid values in influxdb, skip measurement
 			if math.IsNaN(val) || math.IsInf(val, 0) {
-				if debug {
-					log.Printf("Measurement [%s] field [%s] has a NaN or Inf "+
-						"field, skipping",
-						measurement, k)
-				}
+				log.Printf("D! Measurement [%s] field [%s] has a NaN or Inf "+
+					"field, skipping",
+					measurement, k)
 				delete(fields, k)
 				continue
 			}
@@ -135,16 +133,7 @@ func makemetric(
 		}
 	}
 
-	var m telegraf.Metric
-	var err error
-	switch mType {
-	case telegraf.Counter:
-		m, err = telegraf.NewCounterMetric(measurement, tags, fields, t)
-	case telegraf.Gauge:
-		m, err = telegraf.NewGaugeMetric(measurement, tags, fields, t)
-	default:
-		m, err = telegraf.NewMetric(measurement, tags, fields, t)
-	}
+	m, err := metric.New(measurement, tags, fields, t, mType)
 	if err != nil {
 		log.Printf("Error adding point [%s]: %s\n", measurement, err.Error())
 		return nil

internal/models/running_aggregator.go

@@ -4,6 +4,7 @@ import (
 	"time"
 
 	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/metric"
 )
 
 type RunningAggregator struct {
@@ -65,12 +66,13 @@ func (r *RunningAggregator) MakeMetric(
 		nil,
 		r.Config.Filter,
 		false,
-		false,
 		mType,
 		t,
 	)
 
-	m.SetAggregate(true)
+	if m != nil {
+		m.SetAggregate(true)
+	}
 
 	return m
 }
@@ -90,7 +92,7 @@ func (r *RunningAggregator) Add(in telegraf.Metric) bool {
 			return false
 		}
 
-		in, _ = telegraf.NewMetric(name, tags, fields, t)
+		in, _ = metric.New(name, tags, fields, t)
 	}
 
 	r.metrics <- in

internal/models/running_aggregator_test.go

@@ -184,8 +184,8 @@ func TestMakeMetricA(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("RITest value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("RITest value=101i %d", now.UnixNano()),
 	)
 	assert.Equal(
 		t,
@@ -202,8 +202,8 @@ func TestMakeMetricA(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("RITest value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("RITest value=101i %d", now.UnixNano()),
 	)
 	assert.Equal(
 		t,
@@ -220,8 +220,8 @@ func TestMakeMetricA(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("RITest value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("RITest value=101i %d", now.UnixNano()),
 	)
 	assert.Equal(
 		t,

internal/models/running_input.go

@@ -5,15 +5,34 @@ import (
 	"time"
 
 	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/selfstat"
 )
 
+var GlobalMetricsGathered = selfstat.Register("agent", "metrics_gathered", map[string]string{})
+
 type RunningInput struct {
 	Input  telegraf.Input
 	Config *InputConfig
 
 	trace       bool
-	debug       bool
 	defaultTags map[string]string
+
+	MetricsGathered selfstat.Stat
+}
+
+func NewRunningInput(
+	input telegraf.Input,
+	config *InputConfig,
+) *RunningInput {
+	return &RunningInput{
+		Input:  input,
+		Config: config,
+		MetricsGathered: selfstat.Register(
+			"gather",
+			"metrics_gathered",
+			map[string]string{"input": config.Name},
+		),
+	}
 }
 
 // InputConfig containing a name, interval, and filter
@@ -51,26 +70,19 @@ func (r *RunningInput) MakeMetric(
 		r.defaultTags,
 		r.Config.Filter,
 		true,
-		r.debug,
 		mType,
 		t,
 	)
 
 	if r.trace && m != nil {
-		fmt.Println("> " + m.String())
+		fmt.Print("> " + m.String())
 	}
 
+	r.MetricsGathered.Incr(1)
+	GlobalMetricsGathered.Incr(1)
 	return m
 }
 
-func (r *RunningInput) Debug() bool {
-	return r.debug
-}
-
-func (r *RunningInput) SetDebug(debug bool) {
-	r.debug = debug
-}
-
 func (r *RunningInput) Trace() bool {
 	return r.trace
 }

internal/models/running_input_test.go

@@ -13,11 +13,9 @@ import (
 
 func TestMakeMetricNoFields(t *testing.T) {
 	now := time.Now()
-	ri := RunningInput{
-		Config: &InputConfig{
-			Name: "TestRunningInput",
-		},
-	}
+	ri := NewRunningInput(&testInput{}, &InputConfig{
+		Name: "TestRunningInput",
+	})
 
 	m := ri.MakeMetric(
 		"RITest",
@@ -32,11 +30,9 @@ func TestMakeMetricNoFields(t *testing.T) {
 // nil fields should get dropped
 func TestMakeMetricNilFields(t *testing.T) {
 	now := time.Now()
-	ri := RunningInput{
-		Config: &InputConfig{
-			Name: "TestRunningInput",
-		},
-	}
+	ri := NewRunningInput(&testInput{}, &InputConfig{
+		Name: "TestRunningInput",
+	})
 
 	m := ri.MakeMetric(
 		"RITest",
@@ -50,7 +46,7 @@ func TestMakeMetricNilFields(t *testing.T) {
 	)
 	assert.Equal(
 		t,
-		fmt.Sprintf("RITest value=101i %d", now.UnixNano()),
+		fmt.Sprintf("RITest value=101i %d\n", now.UnixNano()),
 		m.String(),
 	)
 }
@@ -58,13 +54,10 @@ func TestMakeMetricNilFields(t *testing.T) {
 // make an untyped, counter, & gauge metric
 func TestMakeMetric(t *testing.T) {
 	now := time.Now()
-	ri := RunningInput{
-		Config: &InputConfig{
-			Name: "TestRunningInput",
-		},
-	}
-	ri.SetDebug(true)
-	assert.Equal(t, true, ri.Debug())
+	ri := NewRunningInput(&testInput{}, &InputConfig{
+		Name: "TestRunningInput",
+	})
+
 	ri.SetTrace(true)
 	assert.Equal(t, true, ri.Trace())
 	assert.Equal(t, "inputs.TestRunningInput", ri.Name())
@@ -78,8 +71,8 @@ func TestMakeMetric(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("RITest value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("RITest value=101i %d", now.UnixNano()),
 	)
 	assert.Equal(
 		t,
@@ -96,8 +89,8 @@ func TestMakeMetric(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("RITest value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("RITest value=101i %d", now.UnixNano()),
 	)
 	assert.Equal(
 		t,
@@ -114,8 +107,8 @@ func TestMakeMetric(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("RITest value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("RITest value=101i %d", now.UnixNano()),
 	)
 	assert.Equal(
 		t,
@@ -126,16 +119,13 @@ func TestMakeMetric(t *testing.T) {
 
 func TestMakeMetricWithPluginTags(t *testing.T) {
 	now := time.Now()
-	ri := RunningInput{
-		Config: &InputConfig{
-			Name: "TestRunningInput",
-			Tags: map[string]string{
-				"foo": "bar",
-			},
+	ri := NewRunningInput(&testInput{}, &InputConfig{
+		Name: "TestRunningInput",
+		Tags: map[string]string{
+			"foo": "bar",
 		},
-	}
-	ri.SetDebug(true)
-	assert.Equal(t, true, ri.Debug())
+	})
+
 	ri.SetTrace(true)
 	assert.Equal(t, true, ri.Trace())
 
@@ -148,24 +138,21 @@ func TestMakeMetricWithPluginTags(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("RITest,foo=bar value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("RITest,foo=bar value=101i %d", now.UnixNano()),
 	)
 }
 
 func TestMakeMetricFilteredOut(t *testing.T) {
 	now := time.Now()
-	ri := RunningInput{
-		Config: &InputConfig{
-			Name: "TestRunningInput",
-			Tags: map[string]string{
-				"foo": "bar",
-			},
-			Filter: Filter{NamePass: []string{"foobar"}},
+	ri := NewRunningInput(&testInput{}, &InputConfig{
+		Name: "TestRunningInput",
+		Tags: map[string]string{
+			"foo": "bar",
 		},
-	}
-	ri.SetDebug(true)
-	assert.Equal(t, true, ri.Debug())
+		Filter: Filter{NamePass: []string{"foobar"}},
+	})
+
 	ri.SetTrace(true)
 	assert.Equal(t, true, ri.Trace())
 	assert.NoError(t, ri.Config.Filter.Compile())
@@ -182,16 +169,13 @@ func TestMakeMetricFilteredOut(t *testing.T) {
 
 func TestMakeMetricWithDaemonTags(t *testing.T) {
 	now := time.Now()
-	ri := RunningInput{
-		Config: &InputConfig{
-			Name: "TestRunningInput",
-		},
-	}
+	ri := NewRunningInput(&testInput{}, &InputConfig{
+		Name: "TestRunningInput",
+	})
 	ri.SetDefaultTags(map[string]string{
 		"foo": "bar",
 	})
-	ri.SetDebug(true)
-	assert.Equal(t, true, ri.Debug())
+
 	ri.SetTrace(true)
 	assert.Equal(t, true, ri.Trace())
 
@@ -204,8 +188,8 @@ func TestMakeMetricWithDaemonTags(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("RITest,foo=bar value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("RITest,foo=bar value=101i %d", now.UnixNano()),
 	)
 }
 
@@ -214,13 +198,10 @@ func TestMakeMetricInfFields(t *testing.T) {
 	inf := math.Inf(1)
 	ninf := math.Inf(-1)
 	now := time.Now()
-	ri := RunningInput{
-		Config: &InputConfig{
-			Name: "TestRunningInput",
-		},
-	}
-	ri.SetDebug(true)
-	assert.Equal(t, true, ri.Debug())
+	ri := NewRunningInput(&testInput{}, &InputConfig{
+		Name: "TestRunningInput",
+	})
+
 	ri.SetTrace(true)
 	assert.Equal(t, true, ri.Trace())
 
@@ -237,20 +218,17 @@ func TestMakeMetricInfFields(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("RITest value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("RITest value=101i %d", now.UnixNano()),
 	)
 }
 
 func TestMakeMetricAllFieldTypes(t *testing.T) {
 	now := time.Now()
-	ri := RunningInput{
-		Config: &InputConfig{
-			Name: "TestRunningInput",
-		},
-	}
-	ri.SetDebug(true)
-	assert.Equal(t, true, ri.Debug())
+	ri := NewRunningInput(&testInput{}, &InputConfig{
+		Name: "TestRunningInput",
+	})
+
 	ri.SetTrace(true)
 	assert.Equal(t, true, ri.Trace())
 
@@ -275,21 +253,28 @@ func TestMakeMetricAllFieldTypes(t *testing.T) {
 		telegraf.Untyped,
 		now,
 	)
-	assert.Equal(
-		t,
-		fmt.Sprintf("RITest a=10i,b=10i,c=10i,d=10i,e=10i,f=10i,g=10i,h=10i,i=10i,j=10,k=9223372036854775807i,l=\"foobar\",m=true %d", now.UnixNano()),
-		m.String(),
-	)
+	assert.Contains(t, m.String(), "a=10i")
+	assert.Contains(t, m.String(), "b=10i")
+	assert.Contains(t, m.String(), "c=10i")
+	assert.Contains(t, m.String(), "d=10i")
+	assert.Contains(t, m.String(), "e=10i")
+	assert.Contains(t, m.String(), "f=10i")
+	assert.Contains(t, m.String(), "g=10i")
+	assert.Contains(t, m.String(), "h=10i")
+	assert.Contains(t, m.String(), "i=10i")
+	assert.Contains(t, m.String(), "j=10")
+	assert.NotContains(t, m.String(), "j=10i")
+	assert.Contains(t, m.String(), "k=9223372036854775807i")
+	assert.Contains(t, m.String(), "l=\"foobar\"")
+	assert.Contains(t, m.String(), "m=true")
 }
 
 func TestMakeMetricNameOverride(t *testing.T) {
 	now := time.Now()
-	ri := RunningInput{
-		Config: &InputConfig{
-			Name:         "TestRunningInput",
-			NameOverride: "foobar",
-		},
-	}
+	ri := NewRunningInput(&testInput{}, &InputConfig{
+		Name:         "TestRunningInput",
+		NameOverride: "foobar",
+	})
 
 	m := ri.MakeMetric(
 		"RITest",
@@ -300,19 +285,17 @@ func TestMakeMetricNameOverride(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("foobar value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("foobar value=101i %d", now.UnixNano()),
 	)
 }
 
 func TestMakeMetricNamePrefix(t *testing.T) {
 	now := time.Now()
-	ri := RunningInput{
-		Config: &InputConfig{
-			Name:              "TestRunningInput",
-			MeasurementPrefix: "foobar_",
-		},
-	}
+	ri := NewRunningInput(&testInput{}, &InputConfig{
+		Name:              "TestRunningInput",
+		MeasurementPrefix: "foobar_",
+	})
 
 	m := ri.MakeMetric(
 		"RITest",
@@ -323,19 +306,17 @@ func TestMakeMetricNamePrefix(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("foobar_RITest value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("foobar_RITest value=101i %d", now.UnixNano()),
 	)
 }
 
 func TestMakeMetricNameSuffix(t *testing.T) {
 	now := time.Now()
-	ri := RunningInput{
-		Config: &InputConfig{
-			Name:              "TestRunningInput",
-			MeasurementSuffix: "_foobar",
-		},
-	}
+	ri := NewRunningInput(&testInput{}, &InputConfig{
+		Name:              "TestRunningInput",
+		MeasurementSuffix: "_foobar",
+	})
 
 	m := ri.MakeMetric(
 		"RITest",
@@ -346,7 +327,13 @@ func TestMakeMetricNameSuffix(t *testing.T) {
 	)
 	assert.Equal(
 		t,
+		fmt.Sprintf("RITest_foobar value=101i %d\n", now.UnixNano()),
 		m.String(),
-		fmt.Sprintf("RITest_foobar value=101i %d", now.UnixNano()),
 	)
 }
+
+type testInput struct{}
+
+func (t *testInput) Description() string                   { return "" }
+func (t *testInput) SampleConfig() string                  { return "" }
+func (t *testInput) Gather(acc telegraf.Accumulator) error { return nil }

internal/models/running_output.go

@@ -6,6 +6,8 @@ import (
 
 	"github.com/influxdata/telegraf"
 	"github.com/influxdata/telegraf/internal/buffer"
+	"github.com/influxdata/telegraf/metric"
+	"github.com/influxdata/telegraf/selfstat"
 )
 
 const (
@@ -21,10 +23,15 @@ type RunningOutput struct {
 	Name              string
 	Output            telegraf.Output
 	Config            *OutputConfig
-	Quiet             bool
 	MetricBufferLimit int
 	MetricBatchSize   int
 
+	MetricsFiltered selfstat.Stat
+	MetricsWritten  selfstat.Stat
+	BufferSize      selfstat.Stat
+	BufferLimit     selfstat.Stat
+	WriteTime       selfstat.Stat
+
 	metrics     *buffer.Buffer
 	failMetrics *buffer.Buffer
 }
@@ -50,29 +57,59 @@ func NewRunningOutput(
 		Config:            conf,
 		MetricBufferLimit: bufferLimit,
 		MetricBatchSize:   batchSize,
+		MetricsWritten: selfstat.Register(
+			"write",
+			"metrics_written",
+			map[string]string{"output": name},
+		),
+		MetricsFiltered: selfstat.Register(
+			"write",
+			"metrics_filtered",
+			map[string]string{"output": name},
+		),
+		BufferSize: selfstat.Register(
+			"write",
+			"buffer_size",
+			map[string]string{"output": name},
+		),
+		BufferLimit: selfstat.Register(
+			"write",
+			"buffer_limit",
+			map[string]string{"output": name},
+		),
+		WriteTime: selfstat.RegisterTiming(
+			"write",
+			"write_time_ns",
+			map[string]string{"output": name},
+		),
 	}
+	ro.BufferLimit.Incr(int64(ro.MetricBufferLimit))
 	return ro
 }
 
 // AddMetric adds a metric to the output. This function can also write cached
 // points if FlushBufferWhenFull is true.
-func (ro *RunningOutput) AddMetric(metric telegraf.Metric) {
+func (ro *RunningOutput) AddMetric(m telegraf.Metric) {
+	if m == nil {
+		return
+	}
 	// Filter any tagexclude/taginclude parameters before adding metric
 	if ro.Config.Filter.IsActive() {
 		// In order to filter out tags, we need to create a new metric, since
 		// metrics are immutable once created.
-		name := metric.Name()
-		tags := metric.Tags()
-		fields := metric.Fields()
-		t := metric.Time()
+		name := m.Name()
+		tags := m.Tags()
+		fields := m.Fields()
+		t := m.Time()
 		if ok := ro.Config.Filter.Apply(name, fields, tags); !ok {
+			ro.MetricsFiltered.Incr(1)
 			return
 		}
 		// error is not possible if creating from another metric, so ignore.
-		metric, _ = telegraf.NewMetric(name, tags, fields, t)
+		m, _ = metric.New(name, tags, fields, t)
 	}
 
-	ro.metrics.Add(metric)
+	ro.metrics.Add(m)
 	if ro.metrics.Len() == ro.MetricBatchSize {
 		batch := ro.metrics.Batch(ro.MetricBatchSize)
 		err := ro.write(batch)
@@ -84,28 +121,21 @@ func (ro *RunningOutput) AddMetric(metric telegraf.Metric) {
 
 // Write writes all cached points to this output.
 func (ro *RunningOutput) Write() error {
-	if !ro.Quiet {
-		log.Printf("I! Output [%s] buffer fullness: %d / %d metrics. "+
-			"Total gathered metrics: %d. Total dropped metrics: %d.",
-			ro.Name,
-			ro.failMetrics.Len()+ro.metrics.Len(),
-			ro.MetricBufferLimit,
-			ro.metrics.Total(),
-			ro.metrics.Drops()+ro.failMetrics.Drops())
-	}
-
+	nFails, nMetrics := ro.failMetrics.Len(), ro.metrics.Len()
+	ro.BufferSize.Set(int64(nFails + nMetrics))
+	log.Printf("D! Output [%s] buffer fullness: %d / %d metrics. ",
+		ro.Name, nFails+nMetrics, ro.MetricBufferLimit)
 	var err error
 	if !ro.failMetrics.IsEmpty() {
-		bufLen := ro.failMetrics.Len()
 		// how many batches of failed writes we need to write.
-		nBatches := bufLen/ro.MetricBatchSize + 1
+		nBatches := nFails/ro.MetricBatchSize + 1
 		batchSize := ro.MetricBatchSize
 
 		for i := 0; i < nBatches; i++ {
 			// If it's the last batch, only grab the metrics that have not had
 			// a write attempt already (this is primarily to preserve order).
 			if i == nBatches-1 {
-				batchSize = bufLen % ro.MetricBatchSize
+				batchSize = nFails % ro.MetricBatchSize
 			}
 			batch := ro.failMetrics.Batch(batchSize)
 			// If we've already failed previous writes, don't bother trying to
@@ -126,6 +156,7 @@ func (ro *RunningOutput) Write() error {
 	if err == nil {
 		err = ro.write(batch)
 	}
+
 	if err != nil {
 		ro.failMetrics.Add(batch...)
 		return err
@@ -134,17 +165,18 @@ func (ro *RunningOutput) Write() error {
 }
 
 func (ro *RunningOutput) write(metrics []telegraf.Metric) error {
-	if metrics == nil || len(metrics) == 0 {
+	nMetrics := len(metrics)
+	if nMetrics == 0 {
 		return nil
 	}
 	start := time.Now()
 	err := ro.Output.Write(metrics)
 	elapsed := time.Since(start)
 	if err == nil {
-		if !ro.Quiet {
-			log.Printf("I! Output [%s] wrote batch of %d metrics in %s\n",
-				ro.Name, len(metrics), elapsed)
-		}
+		log.Printf("D! Output [%s] wrote batch of %d metrics in %s\n",
+			ro.Name, nMetrics, elapsed)
+		ro.MetricsWritten.Incr(int64(nMetrics))
+		ro.WriteTime.Incr(elapsed.Nanoseconds())
 	}
 	return err
 }

internal/models/running_output_test.go

@@ -36,10 +36,9 @@ func BenchmarkRunningOutputAddWrite(b *testing.B) {
 
 	m := &perfOutput{}
 	ro := NewRunningOutput("test", m, conf, 1000, 10000)
-	ro.Quiet = true
 
 	for n := 0; n < b.N; n++ {
-		ro.AddMetric(first5[0])
+		ro.AddMetric(testutil.TestMetric(101, "metric1"))
 		ro.Write()
 	}
 }
@@ -52,10 +51,9 @@ func BenchmarkRunningOutputAddWriteEvery100(b *testing.B) {
 
 	m := &perfOutput{}
 	ro := NewRunningOutput("test", m, conf, 1000, 10000)
-	ro.Quiet = true
 
 	for n := 0; n < b.N; n++ {
-		ro.AddMetric(first5[0])
+		ro.AddMetric(testutil.TestMetric(101, "metric1"))
 		if n%100 == 0 {
 			ro.Write()
 		}
@@ -71,13 +69,29 @@ func BenchmarkRunningOutputAddFailWrites(b *testing.B) {
 	m := &perfOutput{}
 	m.failWrite = true
 	ro := NewRunningOutput("test", m, conf, 1000, 10000)
-	ro.Quiet = true
 
 	for n := 0; n < b.N; n++ {
-		ro.AddMetric(first5[0])
+		ro.AddMetric(testutil.TestMetric(101, "metric1"))
 	}
 }
 
+func TestAddingNilMetric(t *testing.T) {
+	conf := &OutputConfig{
+		Filter: Filter{},
+	}
+
+	m := &mockOutput{}
+	ro := NewRunningOutput("test", m, conf, 1000, 10000)
+
+	ro.AddMetric(nil)
+	ro.AddMetric(nil)
+	ro.AddMetric(nil)
+
+	err := ro.Write()
+	assert.NoError(t, err)
+	assert.Len(t, m.Metrics(), 0)
+}
+
 // Test that NameDrop filters ger properly applied.
 func TestRunningOutput_DropFilter(t *testing.T) {
 	conf := &OutputConfig{
@@ -140,7 +154,7 @@ func TestRunningOutput_TagIncludeNoMatch(t *testing.T) {
 	m := &mockOutput{}
 	ro := NewRunningOutput("test", m, conf, 1000, 10000)
 
-	ro.AddMetric(first5[0])
+	ro.AddMetric(testutil.TestMetric(101, "metric1"))
 	assert.Len(t, m.Metrics(), 0)
 
 	err := ro.Write()
@@ -161,7 +175,7 @@ func TestRunningOutput_TagExcludeMatch(t *testing.T) {
 	m := &mockOutput{}
 	ro := NewRunningOutput("test", m, conf, 1000, 10000)
 
-	ro.AddMetric(first5[0])
+	ro.AddMetric(testutil.TestMetric(101, "metric1"))
 	assert.Len(t, m.Metrics(), 0)
 
 	err := ro.Write()
@@ -182,7 +196,7 @@ func TestRunningOutput_TagExcludeNoMatch(t *testing.T) {
 	m := &mockOutput{}
 	ro := NewRunningOutput("test", m, conf, 1000, 10000)
 
-	ro.AddMetric(first5[0])
+	ro.AddMetric(testutil.TestMetric(101, "metric1"))
 	assert.Len(t, m.Metrics(), 0)
 
 	err := ro.Write()
@@ -203,7 +217,7 @@ func TestRunningOutput_TagIncludeMatch(t *testing.T) {
 	m := &mockOutput{}
 	ro := NewRunningOutput("test", m, conf, 1000, 10000)
 
-	ro.AddMetric(first5[0])
+	ro.AddMetric(testutil.TestMetric(101, "metric1"))
 	assert.Len(t, m.Metrics(), 0)
 
 	err := ro.Write()

logger/logger.go

@@ -4,10 +4,14 @@ import (
 	"io"
 	"log"
 	"os"
+	"regexp"
+	"time"
 
 	"github.com/influxdata/wlog"
 )
 
+var prefixRegex = regexp.MustCompile("^[DIWE]!")
+
 // newTelegrafWriter returns a logging-wrapped writer.
 func newTelegrafWriter(w io.Writer) io.Writer {
 	return &telegrafLog{
@@ -19,8 +23,14 @@ type telegrafLog struct {
 	writer io.Writer
 }
 
-func (t *telegrafLog) Write(p []byte) (n int, err error) {
-	return t.writer.Write(p)
+func (t *telegrafLog) Write(b []byte) (n int, err error) {
+	var line []byte
+	if !prefixRegex.Match(b) {
+		line = append([]byte(time.Now().UTC().Format(time.RFC3339)+" I! "), b...)
+	} else {
+		line = append([]byte(time.Now().UTC().Format(time.RFC3339)+" "), b...)
+	}
+	return t.writer.Write(line)
 }
 
 // SetupLogging configures the logging output.
@@ -30,6 +40,7 @@ func (t *telegrafLog) Write(p []byte) (n int, err error) {
 //           interpreted as stderr. If there is an error opening the file the
 //           logger will fallback to stderr.
 func SetupLogging(debug, quiet bool, logfile string) {
+	log.SetFlags(0)
 	if debug {
 		wlog.SetLevel(wlog.DEBUG)
 	}

logger/logger_test.go

@@ -0,0 +1,75 @@
+package logger
+
+import (
+	"bytes"
+	"io/ioutil"
+	"log"
+	"os"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestWriteLogToFile(t *testing.T) {
+	tmpfile, err := ioutil.TempFile("", "")
+	assert.NoError(t, err)
+	defer func() { os.Remove(tmpfile.Name()) }()
+
+	SetupLogging(false, false, tmpfile.Name())
+	log.Printf("I! TEST")
+	log.Printf("D! TEST") // <- should be ignored
+
+	f, err := ioutil.ReadFile(tmpfile.Name())
+	assert.NoError(t, err)
+	assert.Equal(t, f[19:], []byte("Z I! TEST\n"))
+}
+
+func TestDebugWriteLogToFile(t *testing.T) {
+	tmpfile, err := ioutil.TempFile("", "")
+	assert.NoError(t, err)
+	defer func() { os.Remove(tmpfile.Name()) }()
+
+	SetupLogging(true, false, tmpfile.Name())
+	log.Printf("D! TEST")
+
+	f, err := ioutil.ReadFile(tmpfile.Name())
+	assert.NoError(t, err)
+	assert.Equal(t, f[19:], []byte("Z D! TEST\n"))
+}
+
+func TestErrorWriteLogToFile(t *testing.T) {
+	tmpfile, err := ioutil.TempFile("", "")
+	assert.NoError(t, err)
+	defer func() { os.Remove(tmpfile.Name()) }()
+
+	SetupLogging(false, true, tmpfile.Name())
+	log.Printf("E! TEST")
+	log.Printf("I! TEST") // <- should be ignored
+
+	f, err := ioutil.ReadFile(tmpfile.Name())
+	assert.NoError(t, err)
+	assert.Equal(t, f[19:], []byte("Z E! TEST\n"))
+}
+
+func TestAddDefaultLogLevel(t *testing.T) {
+	tmpfile, err := ioutil.TempFile("", "")
+	assert.NoError(t, err)
+	defer func() { os.Remove(tmpfile.Name()) }()
+
+	SetupLogging(true, false, tmpfile.Name())
+	log.Printf("TEST")
+
+	f, err := ioutil.ReadFile(tmpfile.Name())
+	assert.NoError(t, err)
+	assert.Equal(t, f[19:], []byte("Z I! TEST\n"))
+}
+
+func BenchmarkTelegrafLogWrite(b *testing.B) {
+	var msg = []byte("test")
+	var buf bytes.Buffer
+	w := newTelegrafWriter(&buf)
+	for i := 0; i < b.N; i++ {
+		buf.Reset()
+		w.Write(msg)
+	}
+}

metric.go

@@ -2,9 +2,6 @@ package telegraf
 
 import (
 	"time"
-
-	"github.com/influxdata/influxdb/client/v2"
-	"github.com/influxdata/influxdb/models"
 )
 
 // ValueType is an enumeration of metric types that represent a simple value.
@@ -19,159 +16,47 @@ const (
 )
 
 type Metric interface {
-	// Name returns the measurement name of the metric
+	// Serialize serializes the metric into a line-protocol byte buffer,
+	// including a newline at the end.
+	Serialize() []byte
+	// same as Serialize, but avoids an allocation.
+	// returns number of bytes copied into dst.
+	SerializeTo(dst []byte) int
+	// String is the same as Serialize, but returns a string.
+	String() string
+	// Copy deep-copies the metric.
+	Copy() Metric
+	// Split will attempt to return multiple metrics with the same timestamp
+	// whose string representations are no longer than maxSize.
+	// Metrics with a single field may exceed the requested size.
+	Split(maxSize int) []Metric
+
+	// Tag functions
+	HasTag(key string) bool
+	AddTag(key, value string)
+	RemoveTag(key string)
+
+	// Field functions
+	HasField(key string) bool
+	AddField(key string, value interface{})
+	RemoveField(key string) error
+
+	// Name functions
+	SetName(name string)
+	SetPrefix(prefix string)
+	SetSuffix(suffix string)
+
+	// Getting data structure functions
 	Name() string
-
-	// Name returns the tags associated with the metric
 	Tags() map[string]string
-
-	// Time return the timestamp for the metric
+	Fields() map[string]interface{}
 	Time() time.Time
-
-	// Type returns the metric type. Can be either telegraf.Gauge or telegraf.Counter
-	Type() ValueType
-
-	// UnixNano returns the unix nano time of the metric
 	UnixNano() int64
-
-	// HashID returns a non-cryptographic hash of the metric (name + tags)
-	// NOTE: do not persist & depend on this value to disk.
+	Type() ValueType
+	Len() int // returns the length of the serialized metric, including newline
 	HashID() uint64
 
-	// Fields returns the fields for the metric
-	Fields() map[string]interface{}
-
-	// String returns a line-protocol string of the metric
-	String() string
-
-	// PrecisionString returns a line-protocol string of the metric, at precision
-	PrecisionString(precison string) string
-
-	// Point returns a influxdb client.Point object
-	Point() *client.Point
-
-	// SetAggregate sets the metric's aggregate status
-	// This is so that aggregate metrics don't get re-sent to aggregator plugins
+	// aggregator things:
 	SetAggregate(bool)
-	// IsAggregate returns true if the metric is an aggregate
 	IsAggregate() bool
 }
-
-// metric is a wrapper of the influxdb client.Point struct
-type metric struct {
-	pt models.Point
-
-	mType ValueType
-
-	isaggregate bool
-}
-
-func NewMetricFromPoint(pt models.Point) Metric {
-	return &metric{
-		pt:    pt,
-		mType: Untyped,
-	}
-}
-
-// NewMetric returns an untyped metric.
-func NewMetric(
-	name string,
-	tags map[string]string,
-	fields map[string]interface{},
-	t time.Time,
-) (Metric, error) {
-	pt, err := models.NewPoint(name, models.NewTags(tags), fields, t)
-	if err != nil {
-		return nil, err
-	}
-	return &metric{
-		pt:    pt,
-		mType: Untyped,
-	}, nil
-}
-
-// NewGaugeMetric returns a gauge metric.
-// Gauge metrics should be used when the metric is can arbitrarily go up and
-// down. ie, temperature, memory usage, cpu usage, etc.
-func NewGaugeMetric(
-	name string,
-	tags map[string]string,
-	fields map[string]interface{},
-	t time.Time,
-) (Metric, error) {
-	pt, err := models.NewPoint(name, models.NewTags(tags), fields, t)
-	if err != nil {
-		return nil, err
-	}
-	return &metric{
-		pt:    pt,
-		mType: Gauge,
-	}, nil
-}
-
-// NewCounterMetric returns a Counter metric.
-// Counter metrics should be used when the metric being created is an
-// always-increasing counter. ie, net bytes received, requests served, errors, etc.
-func NewCounterMetric(
-	name string,
-	tags map[string]string,
-	fields map[string]interface{},
-	t time.Time,
-) (Metric, error) {
-	pt, err := models.NewPoint(name, models.NewTags(tags), fields, t)
-	if err != nil {
-		return nil, err
-	}
-	return &metric{
-		pt:    pt,
-		mType: Counter,
-	}, nil
-}
-
-func (m *metric) Name() string {
-	return m.pt.Name()
-}
-
-func (m *metric) Tags() map[string]string {
-	return m.pt.Tags().Map()
-}
-
-func (m *metric) Time() time.Time {
-	return m.pt.Time()
-}
-
-func (m *metric) Type() ValueType {
-	return m.mType
-}
-
-func (m *metric) HashID() uint64 {
-	return m.pt.HashID()
-}
-
-func (m *metric) UnixNano() int64 {
-	return m.pt.UnixNano()
-}
-
-func (m *metric) Fields() map[string]interface{} {
-	return m.pt.Fields()
-}
-
-func (m *metric) String() string {
-	return m.pt.String()
-}
-
-func (m *metric) PrecisionString(precison string) string {
-	return m.pt.PrecisionString(precison)
-}
-
-func (m *metric) Point() *client.Point {
-	return client.NewPointFrom(m.pt)
-}
-
-func (m *metric) IsAggregate() bool {
-	return m.isaggregate
-}
-
-func (m *metric) SetAggregate(b bool) {
-	m.isaggregate = b
-}

metric/escape.go

@@ -0,0 +1,49 @@
+package metric
+
+import (
+	"strings"
+)
+
+var (
+	// escaper is for escaping:
+	//   - tag keys
+	//   - tag values
+	//   - field keys
+	// see https://docs.influxdata.com/influxdb/v1.0/write_protocols/line_protocol_tutorial/#special-characters-and-keywords
+	escaper   = strings.NewReplacer(`,`, `\,`, `"`, `\"`, ` `, `\ `, `=`, `\=`)
+	unEscaper = strings.NewReplacer(`\,`, `,`, `\"`, `"`, `\ `, ` `, `\=`, `=`)
+
+	// nameEscaper is for escaping measurement names only.
+	// see https://docs.influxdata.com/influxdb/v1.0/write_protocols/line_protocol_tutorial/#special-characters-and-keywords
+	nameEscaper   = strings.NewReplacer(`,`, `\,`, ` `, `\ `)
+	nameUnEscaper = strings.NewReplacer(`\,`, `,`, `\ `, ` `)
+
+	// stringFieldEscaper is for escaping string field values only.
+	// see https://docs.influxdata.com/influxdb/v1.0/write_protocols/line_protocol_tutorial/#special-characters-and-keywords
+	stringFieldEscaper   = strings.NewReplacer(`"`, `\"`)
+	stringFieldUnEscaper = strings.NewReplacer(`\"`, `"`)
+)
+
+func escape(s string, t string) string {
+	switch t {
+	case "fieldkey", "tagkey", "tagval":
+		return escaper.Replace(s)
+	case "name":
+		return nameEscaper.Replace(s)
+	case "fieldval":
+		return stringFieldEscaper.Replace(s)
+	}
+	return s
+}
+
+func unescape(s string, t string) string {
+	switch t {
+	case "fieldkey", "tagkey", "tagval":
+		return unEscaper.Replace(s)
+	case "name":
+		return nameUnEscaper.Replace(s)
+	case "fieldval":
+		return stringFieldUnEscaper.Replace(s)
+	}
+	return s
+}

metric/inline_strconv_parse.go

@@ -0,0 +1,38 @@
+package metric
+
+import (
+	"reflect"
+	"strconv"
+	"unsafe"
+)
+
+// parseIntBytes is a zero-alloc wrapper around strconv.ParseInt.
+func parseIntBytes(b []byte, base int, bitSize int) (i int64, err error) {
+	s := unsafeBytesToString(b)
+	return strconv.ParseInt(s, base, bitSize)
+}
+
+// parseFloatBytes is a zero-alloc wrapper around strconv.ParseFloat.
+func parseFloatBytes(b []byte, bitSize int) (float64, error) {
+	s := unsafeBytesToString(b)
+	return strconv.ParseFloat(s, bitSize)
+}
+
+// parseBoolBytes is a zero-alloc wrapper around strconv.ParseBool.
+func parseBoolBytes(b []byte) (bool, error) {
+	return strconv.ParseBool(unsafeBytesToString(b))
+}
+
+// unsafeBytesToString converts a []byte to a string without a heap allocation.
+//
+// It is unsafe, and is intended to prepare input to short-lived functions
+// that require strings.
+func unsafeBytesToString(in []byte) string {
+	src := *(*reflect.SliceHeader)(unsafe.Pointer(&in))
+	dst := reflect.StringHeader{
+		Data: src.Data,
+		Len:  src.Len,
+	}
+	s := *(*string)(unsafe.Pointer(&dst))
+	return s
+}

metric/inline_strconv_parse_test.go

@@ -0,0 +1,103 @@
+package metric
+
+import (
+	"strconv"
+	"testing"
+	"testing/quick"
+)
+
+func TestParseIntBytesEquivalenceFuzz(t *testing.T) {
+	f := func(b []byte, base int, bitSize int) bool {
+		exp, expErr := strconv.ParseInt(string(b), base, bitSize)
+		got, gotErr := parseIntBytes(b, base, bitSize)
+
+		return exp == got && checkErrs(expErr, gotErr)
+	}
+
+	cfg := &quick.Config{
+		MaxCount: 10000,
+	}
+
+	if err := quick.Check(f, cfg); err != nil {
+		t.Fatal(err)
+	}
+}
+
+func TestParseIntBytesValid64bitBase10EquivalenceFuzz(t *testing.T) {
+	buf := []byte{}
+	f := func(n int64) bool {
+		buf = strconv.AppendInt(buf[:0], n, 10)
+
+		exp, expErr := strconv.ParseInt(string(buf), 10, 64)
+		got, gotErr := parseIntBytes(buf, 10, 64)
+
+		return exp == got && checkErrs(expErr, gotErr)
+	}
+
+	cfg := &quick.Config{
+		MaxCount: 10000,
+	}
+
+	if err := quick.Check(f, cfg); err != nil {
+		t.Fatal(err)
+	}
+}
+
+func TestParseFloatBytesEquivalenceFuzz(t *testing.T) {
+	f := func(b []byte, bitSize int) bool {
+		exp, expErr := strconv.ParseFloat(string(b), bitSize)
+		got, gotErr := parseFloatBytes(b, bitSize)
+
+		return exp == got && checkErrs(expErr, gotErr)
+	}
+
+	cfg := &quick.Config{
+		MaxCount: 10000,
+	}
+
+	if err := quick.Check(f, cfg); err != nil {
+		t.Fatal(err)
+	}
+}
+
+func TestParseFloatBytesValid64bitEquivalenceFuzz(t *testing.T) {
+	buf := []byte{}
+	f := func(n float64) bool {
+		buf = strconv.AppendFloat(buf[:0], n, 'f', -1, 64)
+
+		exp, expErr := strconv.ParseFloat(string(buf), 64)
+		got, gotErr := parseFloatBytes(buf, 64)
+
+		return exp == got && checkErrs(expErr, gotErr)
+	}
+
+	cfg := &quick.Config{
+		MaxCount: 10000,
+	}
+
+	if err := quick.Check(f, cfg); err != nil {
+		t.Fatal(err)
+	}
+}
+
+func TestParseBoolBytesEquivalence(t *testing.T) {
+	var buf []byte
+	for _, s := range []string{"1", "t", "T", "TRUE", "true", "True", "0", "f", "F", "FALSE", "false", "False", "fail", "TrUe", "FAlSE", "numbers", ""} {
+		buf = append(buf[:0], s...)
+
+		exp, expErr := strconv.ParseBool(s)
+		got, gotErr := parseBoolBytes(buf)
+
+		if got != exp || !checkErrs(expErr, gotErr) {
+			t.Errorf("Failed to parse boolean value %q correctly: wanted (%t, %v), got (%t, %v)", s, exp, expErr, got, gotErr)
+		}
+	}
+}
+
+func checkErrs(a, b error) bool {
+	if (a == nil) != (b == nil) {
+		return false
+	}
+
+	return a == nil || a.Error() == b.Error()
+}

metric/metric.go

@@ -0,0 +1,585 @@
+package metric
+
+import (
+	"bytes"
+	"fmt"
+	"hash/fnv"
+	"sort"
+	"strconv"
+	"time"
+
+	"github.com/influxdata/telegraf"
+)
+
+const MaxInt = int(^uint(0) >> 1)
+
+func New(
+	name string,
+	tags map[string]string,
+	fields map[string]interface{},
+	t time.Time,
+	mType ...telegraf.ValueType,
+) (telegraf.Metric, error) {
+	if len(fields) == 0 {
+		return nil, fmt.Errorf("Metric cannot be made without any fields")
+	}
+	if len(name) == 0 {
+		return nil, fmt.Errorf("Metric cannot be made with an empty name")
+	}
+
+	var thisType telegraf.ValueType
+	if len(mType) > 0 {
+		thisType = mType[0]
+	} else {
+		thisType = telegraf.Untyped
+	}
+
+	m := &metric{
+		name:  []byte(escape(name, "name")),
+		t:     []byte(fmt.Sprint(t.UnixNano())),
+		nsec:  t.UnixNano(),
+		mType: thisType,
+	}
+
+	// pre-allocate exact size of the tags slice
+	taglen := 0
+	for k, v := range tags {
+		if len(k) == 0 || len(v) == 0 {
+			continue
+		}
+		taglen += 2 + len(escape(k, "tagkey")) + len(escape(v, "tagval"))
+	}
+	m.tags = make([]byte, taglen)
+
+	i := 0
+	for k, v := range tags {
+		if len(k) == 0 || len(v) == 0 {
+			continue
+		}
+		m.tags[i] = ','
+		i++
+		i += copy(m.tags[i:], escape(k, "tagkey"))
+		m.tags[i] = '='
+		i++
+		i += copy(m.tags[i:], escape(v, "tagval"))
+	}
+
+	// pre-allocate capacity of the fields slice
+	fieldlen := 0
+	for k, _ := range fields {
+		// 10 bytes is completely arbitrary, but will at least prevent some
+		// amount of allocations. There's a small possibility this will create
+		// slightly more allocations for a metric that has many short fields.
+		fieldlen += len(k) + 10
+	}
+	m.fields = make([]byte, 0, fieldlen)
+
+	i = 0
+	for k, v := range fields {
+		if i != 0 {
+			m.fields = append(m.fields, ',')
+		}
+		m.fields = appendField(m.fields, k, v)
+		i++
+	}
+
+	return m, nil
+}
+
+// indexUnescapedByte finds the index of the first byte equal to b in buf that
+// is not escaped. Returns -1 if not found.
+func indexUnescapedByte(buf []byte, b byte) int {
+	var keyi int
+	for {
+		i := bytes.IndexByte(buf[keyi:], b)
+		if i == -1 {
+			return -1
+		} else if i == 0 {
+			break
+		}
+		keyi += i
+		if countBackslashes(buf, keyi-1)%2 == 0 {
+			break
+		} else {
+			keyi++
+		}
+	}
+	return keyi
+}
+
+// countBackslashes counts the number of preceding backslashes starting at
+// the 'start' index.
+func countBackslashes(buf []byte, index int) int {
+	var count int
+	for {
+		if index < 0 {
+			return count
+		}
+		if buf[index] == '\\' {
+			count++
+			index--
+		} else {
+			break
+		}
+	}
+	return count
+}
+
+type metric struct {
+	name   []byte
+	tags   []byte
+	fields []byte
+	t      []byte
+
+	mType     telegraf.ValueType
+	aggregate bool
+
+	// cached values for reuse in "get" functions
+	hashID uint64
+	nsec   int64
+}
+
+func (m *metric) String() string {
+	return string(m.name) + string(m.tags) + " " + string(m.fields) + " " + string(m.t) + "\n"
+}
+
+func (m *metric) SetAggregate(b bool) {
+	m.aggregate = b
+}
+
+func (m *metric) IsAggregate() bool {
+	return m.aggregate
+}
+
+func (m *metric) Type() telegraf.ValueType {
+	return m.mType
+}
+
+func (m *metric) Len() int {
+	// 3 is for 2 spaces surrounding the fields array + newline at the end.
+	return len(m.name) + len(m.tags) + len(m.fields) + len(m.t) + 3
+}
+
+func (m *metric) Serialize() []byte {
+	tmp := make([]byte, m.Len())
+	i := 0
+	i += copy(tmp[i:], m.name)
+	i += copy(tmp[i:], m.tags)
+	tmp[i] = ' '
+	i++
+	i += copy(tmp[i:], m.fields)
+	tmp[i] = ' '
+	i++
+	i += copy(tmp[i:], m.t)
+	tmp[i] = '\n'
+	return tmp
+}
+
+func (m *metric) SerializeTo(dst []byte) int {
+	i := 0
+	if i >= len(dst) {
+		return i
+	}
+
+	i += copy(dst[i:], m.name)
+	if i >= len(dst) {
+		return i
+	}
+
+	i += copy(dst[i:], m.tags)
+	if i >= len(dst) {
+		return i
+	}
+
+	dst[i] = ' '
+	i++
+	if i >= len(dst) {
+		return i
+	}
+
+	i += copy(dst[i:], m.fields)
+	if i >= len(dst) {
+		return i
+	}
+
+	dst[i] = ' '
+	i++
+	if i >= len(dst) {
+		return i
+	}
+
+	i += copy(dst[i:], m.t)
+	if i >= len(dst) {
+		return i
+	}
+	dst[i] = '\n'
+
+	return i + 1
+}
+
+func (m *metric) Split(maxSize int) []telegraf.Metric {
+	if m.Len() < maxSize {
+		return []telegraf.Metric{m}
+	}
+	var out []telegraf.Metric
+
+	// constant number of bytes for each metric (in addition to field bytes)
+	constant := len(m.name) + len(m.tags) + len(m.t) + 3
+	// currently selected fields
+	fields := make([]byte, 0, maxSize)
+
+	i := 0
+	for {
+		if i >= len(m.fields) {
+			// hit the end of the field byte slice
+			if len(fields) > 0 {
+				out = append(out, copyWith(m.name, m.tags, fields, m.t))
+			}
+			break
+		}
+
+		// find the end of the next field
+		j := indexUnescapedByte(m.fields[i:], ',')
+		if j == -1 {
+			j = len(m.fields)
+		} else {
+			j += i
+		}
+
+		// if true, then we need to create a metric _not_ including the currently
+		// selected field
+		if len(m.fields[i:j])+len(fields)+constant > maxSize {
+			// if false, then we'll create a metric including the currently
+			// selected field anyways. This means that the given maxSize is too
+			// small for a single field to fit.
+			if len(fields) > 0 {
+				out = append(out, copyWith(m.name, m.tags, fields, m.t))
+			}
+
+			fields = make([]byte, 0, maxSize)
+		}
+		if len(fields) > 0 {
+			fields = append(fields, ',')
+		}
+		fields = append(fields, m.fields[i:j]...)
+
+		i = j + 1
+	}
+	return out
+}
+
+func (m *metric) Fields() map[string]interface{} {
+	fieldMap := map[string]interface{}{}
+	i := 0
+	for {
+		if i >= len(m.fields) {
+			break
+		}
+		// end index of field key
+		i1 := indexUnescapedByte(m.fields[i:], '=')
+		if i1 == -1 {
+			break
+		}
+		// start index of field value
+		i2 := i1 + 1
+
+		// end index of field value
+		var i3 int
+		if m.fields[i:][i2] == '"' {
+			i3 = indexUnescapedByte(m.fields[i:][i2+1:], '"')
+			if i3 == -1 {
+				i3 = len(m.fields[i:])
+			}
+			i3 += i2 + 2 // increment index to the comma
+		} else {
+			i3 = indexUnescapedByte(m.fields[i:], ',')
+			if i3 == -1 {
+				i3 = len(m.fields[i:])
+			}
+		}
+
+		switch m.fields[i:][i2] {
+		case '"':
+			// string field
+			fieldMap[unescape(string(m.fields[i:][0:i1]), "fieldkey")] = unescape(string(m.fields[i:][i2+1:i3-1]), "fieldval")
+		case '-', '0', '1', '2', '3', '4', '5', '6', '7', '8', '9':
+			// number field
+			switch m.fields[i:][i3-1] {
+			case 'i':
+				// integer field
+				n, err := parseIntBytes(m.fields[i:][i2:i3-1], 10, 64)
+				if err == nil {
+					fieldMap[unescape(string(m.fields[i:][0:i1]), "fieldkey")] = n
+				} else {
+					// TODO handle error or just ignore field silently?
+				}
+			default:
+				// float field
+				n, err := parseFloatBytes(m.fields[i:][i2:i3], 64)
+				if err == nil {
+					fieldMap[unescape(string(m.fields[i:][0:i1]), "fieldkey")] = n
+				} else {
+					// TODO handle error or just ignore field silently?
+				}
+			}
+		case 'T', 't':
+			fieldMap[unescape(string(m.fields[i:][0:i1]), "fieldkey")] = true
+		case 'F', 'f':
+			fieldMap[unescape(string(m.fields[i:][0:i1]), "fieldkey")] = false
+		default:
+			// TODO handle unsupported field type
+		}
+
+		i += i3 + 1
+	}
+
+	return fieldMap
+}
+
+func (m *metric) Tags() map[string]string {
+	tagMap := map[string]string{}
+	if len(m.tags) == 0 {
+		return tagMap
+	}
+
+	i := 0
+	for {
+		// start index of tag key
+		i0 := indexUnescapedByte(m.tags[i:], ',') + 1
+		if i0 == 0 {
+			// didn't find a tag start
+			break
+		}
+		// end index of tag key
+		i1 := indexUnescapedByte(m.tags[i:], '=')
+		// start index of tag value
+		i2 := i1 + 1
+		// end index of tag value (starting from i2)
+		i3 := indexUnescapedByte(m.tags[i+i2:], ',')
+		if i3 == -1 {
+			tagMap[unescape(string(m.tags[i:][i0:i1]), "tagkey")] = unescape(string(m.tags[i:][i2:]), "tagval")
+			break
+		}
+		tagMap[unescape(string(m.tags[i:][i0:i1]), "tagkey")] = unescape(string(m.tags[i:][i2:i2+i3]), "tagval")
+		// increment start index for the next tag
+		i += i2 + i3
+	}
+
+	return tagMap
+}
+
+func (m *metric) Name() string {
+	return unescape(string(m.name), "name")
+}
+
+func (m *metric) Time() time.Time {
+	// assume metric has been verified already and ignore error:
+	if m.nsec == 0 {
+		m.nsec, _ = parseIntBytes(m.t, 10, 64)
+	}
+	return time.Unix(0, m.nsec)
+}
+
+func (m *metric) UnixNano() int64 {
+	// assume metric has been verified already and ignore error:
+	if m.nsec == 0 {
+		m.nsec, _ = parseIntBytes(m.t, 10, 64)
+	}
+	return m.nsec
+}
+
+func (m *metric) SetName(name string) {
+	m.hashID = 0
+	m.name = []byte(nameEscaper.Replace(name))
+}
+
+func (m *metric) SetPrefix(prefix string) {
+	m.hashID = 0
+	m.name = append([]byte(nameEscaper.Replace(prefix)), m.name...)
+}
+
+func (m *metric) SetSuffix(suffix string) {
+	m.hashID = 0
+	m.name = append(m.name, []byte(nameEscaper.Replace(suffix))...)
+}
+
+func (m *metric) AddTag(key, value string) {
+	m.RemoveTag(key)
+	m.tags = append(m.tags, []byte(","+escape(key, "tagkey")+"="+escape(value, "tagval"))...)
+}
+
+func (m *metric) HasTag(key string) bool {
+	i := bytes.Index(m.tags, []byte(escape(key, "tagkey")+"="))
+	if i == -1 {
+		return false
+	}
+	return true
+}
+
+func (m *metric) RemoveTag(key string) {
+	m.hashID = 0
+
+	i := bytes.Index(m.tags, []byte(escape(key, "tagkey")+"="))
+	if i == -1 {
+		return
+	}
+
+	tmp := m.tags[0 : i-1]
+	j := indexUnescapedByte(m.tags[i:], ',')
+	if j != -1 {
+		tmp = append(tmp, m.tags[i+j:]...)
+	}
+	m.tags = tmp
+	return
+}
+
+func (m *metric) AddField(key string, value interface{}) {
+	m.fields = append(m.fields, ',')
+	m.fields = appendField(m.fields, key, value)
+}
+
+func (m *metric) HasField(key string) bool {
+	i := bytes.Index(m.fields, []byte(escape(key, "tagkey")+"="))
+	if i == -1 {
+		return false
+	}
+	return true
+}
+
+func (m *metric) RemoveField(key string) error {
+	i := bytes.Index(m.fields, []byte(escape(key, "tagkey")+"="))
+	if i == -1 {
+		return nil
+	}
+
+	var tmp []byte
+	if i != 0 {
+		tmp = m.fields[0 : i-1]
+	}
+	j := indexUnescapedByte(m.fields[i:], ',')
+	if j != -1 {
+		tmp = append(tmp, m.fields[i+j:]...)
+	}
+
+	if len(tmp) == 0 {
+		return fmt.Errorf("Metric cannot remove final field: %s", m.fields)
+	}
+
+	m.fields = tmp
+	return nil
+}
+
+func (m *metric) Copy() telegraf.Metric {
+	return copyWith(m.name, m.tags, m.fields, m.t)
+}
+
+func copyWith(name, tags, fields, t []byte) telegraf.Metric {
+	out := metric{
+		name:   make([]byte, len(name)),
+		tags:   make([]byte, len(tags)),
+		fields: make([]byte, len(fields)),
+		t:      make([]byte, len(t)),
+	}
+	copy(out.name, name)
+	copy(out.tags, tags)
+	copy(out.fields, fields)
+	copy(out.t, t)
+	return &out
+}
+
+func (m *metric) HashID() uint64 {
+	if m.hashID == 0 {
+		h := fnv.New64a()
+		h.Write(m.name)
+
+		tags := m.Tags()
+		tmp := make([]string, len(tags))
+		i := 0
+		for k, v := range tags {
+			tmp[i] = k + v
+			i++
+		}
+		sort.Strings(tmp)
+
+		for _, s := range tmp {
+			h.Write([]byte(s))
+		}
+
+		m.hashID = h.Sum64()
+	}
+	return m.hashID
+}
+
+func appendField(b []byte, k string, v interface{}) []byte {
+	if v == nil {
+		return b
+	}
+	b = append(b, []byte(escape(k, "tagkey")+"=")...)
+
+	// check popular types first
+	switch v := v.(type) {
+	case float64:
+		b = strconv.AppendFloat(b, v, 'f', -1, 64)
+	case int64:
+		b = strconv.AppendInt(b, v, 10)
+		b = append(b, 'i')
+	case string:
+		b = append(b, '"')
+		b = append(b, []byte(escape(v, "fieldval"))...)
+		b = append(b, '"')
+	case bool:
+		b = strconv.AppendBool(b, v)
+	case int32:
+		b = strconv.AppendInt(b, int64(v), 10)
+		b = append(b, 'i')
+	case int16:
+		b = strconv.AppendInt(b, int64(v), 10)
+		b = append(b, 'i')
+	case int8:
+		b = strconv.AppendInt(b, int64(v), 10)
+		b = append(b, 'i')
+	case int:
+		b = strconv.AppendInt(b, int64(v), 10)
+		b = append(b, 'i')
+	case uint64:
+		// Cap uints above the maximum int value
+		var intv int64
+		if v <= uint64(MaxInt) {
+			intv = int64(v)
+		} else {
+			intv = int64(MaxInt)
+		}
+		b = strconv.AppendInt(b, intv, 10)
+		b = append(b, 'i')
+	case uint32:
+		b = strconv.AppendInt(b, int64(v), 10)
+		b = append(b, 'i')
+	case uint16:
+		b = strconv.AppendInt(b, int64(v), 10)
+		b = append(b, 'i')
+	case uint8:
+		b = strconv.AppendInt(b, int64(v), 10)
+		b = append(b, 'i')
+	case uint:
+		// Cap uints above the maximum int value
+		var intv int64
+		if v <= uint(MaxInt) {
+			intv = int64(v)
+		} else {
+			intv = int64(MaxInt)
+		}
+		b = strconv.AppendInt(b, intv, 10)
+		b = append(b, 'i')
+	case float32:
+		b = strconv.AppendFloat(b, float64(v), 'f', -1, 32)
+	case []byte:
+		b = append(b, v...)
+	default:
+		// Can't determine the type, so convert to string
+		b = append(b, '"')
+		b = append(b, []byte(escape(fmt.Sprintf("%v", v), "fieldval"))...)
+		b = append(b, '"')
+	}
+
+	return b
+}

metric/metric_benchmark_test.go

@@ -0,0 +1,148 @@
+package metric
+
+import (
+	"fmt"
+	"testing"
+	"time"
+
+	"github.com/influxdata/telegraf"
+)
+
+// vars for making sure that the compiler doesnt optimize out the benchmarks:
+var (
+	s      string
+	I      interface{}
+	tags   map[string]string
+	fields map[string]interface{}
+)
+
+func BenchmarkNewMetric(b *testing.B) {
+	var mt telegraf.Metric
+	for n := 0; n < b.N; n++ {
+		mt, _ = New("test_metric",
+			map[string]string{
+				"test_tag_1": "tag_value_1",
+				"test_tag_2": "tag_value_2",
+				"test_tag_3": "tag_value_3",
+			},
+			map[string]interface{}{
+				"string_field": "string",
+				"int_field":    int64(1000),
+				"float_field":  float64(2.1),
+			},
+			time.Now(),
+		)
+	}
+	s = string(mt.String())
+}
+
+func BenchmarkAddTag(b *testing.B) {
+	var mt telegraf.Metric
+	mt = &metric{
+		name:   []byte("cpu"),
+		tags:   []byte(",host=localhost"),
+		fields: []byte("a=101"),
+		t:      []byte("1480614053000000000"),
+	}
+	for n := 0; n < b.N; n++ {
+		mt.AddTag("foo", "bar")
+	}
+	s = string(mt.String())
+}
+
+func BenchmarkSplit(b *testing.B) {
+	var mt telegraf.Metric
+	mt = &metric{
+		name:   []byte("cpu"),
+		tags:   []byte(",host=localhost"),
+		fields: []byte("a=101,b=10i,c=10101,d=101010,e=42"),
+		t:      []byte("1480614053000000000"),
+	}
+	var metrics []telegraf.Metric
+	for n := 0; n < b.N; n++ {
+		metrics = mt.Split(60)
+	}
+	s = string(metrics[0].String())
+}
+
+func BenchmarkTags(b *testing.B) {
+	for n := 0; n < b.N; n++ {
+		var mt, _ = New("test_metric",
+			map[string]string{
+				"test_tag_1": "tag_value_1",
+				"test_tag_2": "tag_value_2",
+				"test_tag_3": "tag_value_3",
+			},
+			map[string]interface{}{
+				"string_field": "string",
+				"int_field":    int64(1000),
+				"float_field":  float64(2.1),
+			},
+			time.Now(),
+		)
+		tags = mt.Tags()
+	}
+	s = fmt.Sprint(tags)
+}
+
+func BenchmarkFields(b *testing.B) {
+	for n := 0; n < b.N; n++ {
+		var mt, _ = New("test_metric",
+			map[string]string{
+				"test_tag_1": "tag_value_1",
+				"test_tag_2": "tag_value_2",
+				"test_tag_3": "tag_value_3",
+			},
+			map[string]interface{}{
+				"string_field": "string",
+				"int_field":    int64(1000),
+				"float_field":  float64(2.1),
+			},
+			time.Now(),
+		)
+		fields = mt.Fields()
+	}
+	s = fmt.Sprint(fields)
+}
+
+func BenchmarkString(b *testing.B) {
+	mt, _ := New("test_metric",
+		map[string]string{
+			"test_tag_1": "tag_value_1",
+			"test_tag_2": "tag_value_2",
+			"test_tag_3": "tag_value_3",
+		},
+		map[string]interface{}{
+			"string_field": "string",
+			"int_field":    int64(1000),
+			"float_field":  float64(2.1),
+		},
+		time.Now(),
+	)
+	var S string
+	for n := 0; n < b.N; n++ {
+		S = mt.String()
+	}
+	s = S
+}
+
+func BenchmarkSerialize(b *testing.B) {
+	mt, _ := New("test_metric",
+		map[string]string{
+			"test_tag_1": "tag_value_1",
+			"test_tag_2": "tag_value_2",
+			"test_tag_3": "tag_value_3",
+		},
+		map[string]interface{}{
+			"string_field": "string",
+			"int_field":    int64(1000),
+			"float_field":  float64(2.1),
+		},
+		time.Now(),
+	)
+	var B []byte
+	for n := 0; n < b.N; n++ {
+		B = mt.Serialize()
+	}
+	s = string(B)
+}

metric/metric_test.go

@@ -0,0 +1,650 @@
+package metric
+
+import (
+	"fmt"
+	"math"
+	"regexp"
+	"testing"
+	"time"
+
+	"github.com/influxdata/telegraf"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestNewMetric(t *testing.T) {
+	now := time.Now()
+
+	tags := map[string]string{
+		"host":       "localhost",
+		"datacenter": "us-east-1",
+	}
+	fields := map[string]interface{}{
+		"usage_idle": float64(99),
+		"usage_busy": float64(1),
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	assert.Equal(t, telegraf.Untyped, m.Type())
+	assert.Equal(t, tags, m.Tags())
+	assert.Equal(t, fields, m.Fields())
+	assert.Equal(t, "cpu", m.Name())
+	assert.Equal(t, now, m.Time())
+	assert.Equal(t, now.UnixNano(), m.UnixNano())
+}
+
+func TestNewErrors(t *testing.T) {
+	// creating a metric with an empty name produces an error:
+	m, err := New(
+		"",
+		map[string]string{
+			"datacenter": "us-east-1",
+			"mytag":      "foo",
+			"another":    "tag",
+		},
+		map[string]interface{}{
+			"value": float64(1),
+		},
+		time.Now(),
+	)
+	assert.Error(t, err)
+	assert.Nil(t, m)
+
+	// creating a metric with empty fields produces an error:
+	m, err = New(
+		"foobar",
+		map[string]string{
+			"datacenter": "us-east-1",
+			"mytag":      "foo",
+			"another":    "tag",
+		},
+		map[string]interface{}{},
+		time.Now(),
+	)
+	assert.Error(t, err)
+	assert.Nil(t, m)
+}
+
+func TestNewMetric_Tags(t *testing.T) {
+	now := time.Now()
+	tags := map[string]string{
+		"host":       "localhost",
+		"datacenter": "us-east-1",
+	}
+	fields := map[string]interface{}{
+		"value": float64(1),
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	assert.True(t, m.HasTag("host"))
+	assert.True(t, m.HasTag("datacenter"))
+
+	m.AddTag("newtag", "foo")
+	assert.True(t, m.HasTag("newtag"))
+
+	m.RemoveTag("host")
+	assert.False(t, m.HasTag("host"))
+	assert.True(t, m.HasTag("newtag"))
+	assert.True(t, m.HasTag("datacenter"))
+
+	m.RemoveTag("datacenter")
+	assert.False(t, m.HasTag("datacenter"))
+	assert.True(t, m.HasTag("newtag"))
+	assert.Equal(t, map[string]string{"newtag": "foo"}, m.Tags())
+
+	m.RemoveTag("newtag")
+	assert.False(t, m.HasTag("newtag"))
+	assert.Equal(t, map[string]string{}, m.Tags())
+
+	assert.Equal(t, "cpu value=1 "+fmt.Sprint(now.UnixNano())+"\n", m.String())
+}
+
+func TestSerialize(t *testing.T) {
+	now := time.Now()
+	tags := map[string]string{
+		"datacenter": "us-east-1",
+	}
+	fields := map[string]interface{}{
+		"value": float64(1),
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	assert.Equal(t,
+		[]byte("cpu,datacenter=us-east-1 value=1 "+fmt.Sprint(now.UnixNano())+"\n"),
+		m.Serialize())
+
+	m.RemoveTag("datacenter")
+	assert.Equal(t,
+		[]byte("cpu value=1 "+fmt.Sprint(now.UnixNano())+"\n"),
+		m.Serialize())
+}
+
+func TestHashID(t *testing.T) {
+	m, _ := New(
+		"cpu",
+		map[string]string{
+			"datacenter": "us-east-1",
+			"mytag":      "foo",
+			"another":    "tag",
+		},
+		map[string]interface{}{
+			"value": float64(1),
+		},
+		time.Now(),
+	)
+	hash := m.HashID()
+
+	// adding a field doesn't change the hash:
+	m.AddField("foo", int64(100))
+	assert.Equal(t, hash, m.HashID())
+
+	// removing a non-existent tag doesn't change the hash:
+	m.RemoveTag("no-op")
+	assert.Equal(t, hash, m.HashID())
+
+	// adding a tag does change it:
+	m.AddTag("foo", "bar")
+	assert.NotEqual(t, hash, m.HashID())
+	hash = m.HashID()
+
+	// removing a tag also changes it:
+	m.RemoveTag("mytag")
+	assert.NotEqual(t, hash, m.HashID())
+}
+
+func TestHashID_Consistency(t *testing.T) {
+	m, _ := New(
+		"cpu",
+		map[string]string{
+			"datacenter": "us-east-1",
+			"mytag":      "foo",
+			"another":    "tag",
+		},
+		map[string]interface{}{
+			"value": float64(1),
+		},
+		time.Now(),
+	)
+	hash := m.HashID()
+
+	for i := 0; i < 1000; i++ {
+		m2, _ := New(
+			"cpu",
+			map[string]string{
+				"datacenter": "us-east-1",
+				"mytag":      "foo",
+				"another":    "tag",
+			},
+			map[string]interface{}{
+				"value": float64(1),
+			},
+			time.Now(),
+		)
+		assert.Equal(t, hash, m2.HashID())
+	}
+}
+
+func TestNewMetric_NameModifiers(t *testing.T) {
+	now := time.Now()
+	tags := map[string]string{}
+	fields := map[string]interface{}{
+		"value": float64(1),
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	hash := m.HashID()
+	suffix := fmt.Sprintf(" value=1 %d\n", now.UnixNano())
+	assert.Equal(t, "cpu"+suffix, m.String())
+
+	m.SetPrefix("pre_")
+	assert.NotEqual(t, hash, m.HashID())
+	hash = m.HashID()
+	assert.Equal(t, "pre_cpu"+suffix, m.String())
+
+	m.SetSuffix("_post")
+	assert.NotEqual(t, hash, m.HashID())
+	hash = m.HashID()
+	assert.Equal(t, "pre_cpu_post"+suffix, m.String())
+
+	m.SetName("mem")
+	assert.NotEqual(t, hash, m.HashID())
+	assert.Equal(t, "mem"+suffix, m.String())
+}
+
+func TestNewMetric_FieldModifiers(t *testing.T) {
+	now := time.Now()
+	tags := map[string]string{
+		"host": "localhost",
+	}
+	fields := map[string]interface{}{
+		"value": float64(1),
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	assert.True(t, m.HasField("value"))
+	assert.False(t, m.HasField("foo"))
+
+	m.AddField("newfield", "foo")
+	assert.True(t, m.HasField("newfield"))
+
+	assert.NoError(t, m.RemoveField("newfield"))
+	assert.False(t, m.HasField("newfield"))
+
+	// don't allow user to remove all fields:
+	assert.Error(t, m.RemoveField("value"))
+
+	m.AddField("value2", int64(101))
+	assert.NoError(t, m.RemoveField("value"))
+	assert.False(t, m.HasField("value"))
+}
+
+func TestNewMetric_Fields(t *testing.T) {
+	now := time.Now()
+	tags := map[string]string{
+		"host": "localhost",
+	}
+	fields := map[string]interface{}{
+		"float":  float64(1),
+		"int":    int64(1),
+		"bool":   true,
+		"false":  false,
+		"string": "test",
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	assert.Equal(t, fields, m.Fields())
+}
+
+func TestNewMetric_Time(t *testing.T) {
+	now := time.Now()
+	tags := map[string]string{
+		"host": "localhost",
+	}
+	fields := map[string]interface{}{
+		"float":  float64(1),
+		"int":    int64(1),
+		"bool":   true,
+		"false":  false,
+		"string": "test",
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+	m = m.Copy()
+	m2 := m.Copy()
+
+	assert.Equal(t, now.UnixNano(), m.Time().UnixNano())
+	assert.Equal(t, now.UnixNano(), m2.UnixNano())
+}
+
+func TestNewMetric_Copy(t *testing.T) {
+	now := time.Now()
+	tags := map[string]string{}
+	fields := map[string]interface{}{
+		"float": float64(1),
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+	m2 := m.Copy()
+
+	assert.Equal(t,
+		fmt.Sprintf("cpu float=1 %d\n", now.UnixNano()),
+		m.String())
+	m.AddTag("host", "localhost")
+	assert.Equal(t,
+		fmt.Sprintf("cpu,host=localhost float=1 %d\n", now.UnixNano()),
+		m.String())
+
+	assert.Equal(t,
+		fmt.Sprintf("cpu float=1 %d\n", now.UnixNano()),
+		m2.String())
+}
+
+func TestNewMetric_AllTypes(t *testing.T) {
+	now := time.Now()
+	tags := map[string]string{}
+	fields := map[string]interface{}{
+		"float64":     float64(1),
+		"float32":     float32(1),
+		"int64":       int64(1),
+		"int32":       int32(1),
+		"int16":       int16(1),
+		"int8":        int8(1),
+		"int":         int(1),
+		"uint64":      uint64(1),
+		"uint32":      uint32(1),
+		"uint16":      uint16(1),
+		"uint8":       uint8(1),
+		"uint":        uint(1),
+		"bytes":       []byte("foo"),
+		"nil":         nil,
+		"maxuint64":   uint64(MaxInt) + 10,
+		"maxuint":     uint(MaxInt) + 10,
+		"unsupported": []int{1, 2},
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	assert.Contains(t, m.String(), "float64=1")
+	assert.Contains(t, m.String(), "float32=1")
+	assert.Contains(t, m.String(), "int64=1i")
+	assert.Contains(t, m.String(), "int32=1i")
+	assert.Contains(t, m.String(), "int16=1i")
+	assert.Contains(t, m.String(), "int8=1i")
+	assert.Contains(t, m.String(), "int=1i")
+	assert.Contains(t, m.String(), "uint64=1i")
+	assert.Contains(t, m.String(), "uint32=1i")
+	assert.Contains(t, m.String(), "uint16=1i")
+	assert.Contains(t, m.String(), "uint8=1i")
+	assert.Contains(t, m.String(), "uint=1i")
+	assert.NotContains(t, m.String(), "nil")
+	assert.Contains(t, m.String(), fmt.Sprintf("maxuint64=%di", MaxInt))
+	assert.Contains(t, m.String(), fmt.Sprintf("maxuint=%di", MaxInt))
+}
+
+func TestIndexUnescapedByte(t *testing.T) {
+	tests := []struct {
+		in       []byte
+		b        byte
+		expected int
+	}{
+		{
+			in:       []byte(`foobar`),
+			b:        'b',
+			expected: 3,
+		},
+		{
+			in:       []byte(`foo\bar`),
+			b:        'b',
+			expected: -1,
+		},
+		{
+			in:       []byte(`foo\\bar`),
+			b:        'b',
+			expected: 5,
+		},
+		{
+			in:       []byte(`foobar`),
+			b:        'f',
+			expected: 0,
+		},
+		{
+			in:       []byte(`foobar`),
+			b:        'r',
+			expected: 5,
+		},
+		{
+			in:       []byte(`\foobar`),
+			b:        'f',
+			expected: -1,
+		},
+	}
+
+	for _, test := range tests {
+		got := indexUnescapedByte(test.in, test.b)
+		assert.Equal(t, test.expected, got)
+	}
+}
+
+func TestNewGaugeMetric(t *testing.T) {
+	now := time.Now()
+
+	tags := map[string]string{
+		"host":       "localhost",
+		"datacenter": "us-east-1",
+	}
+	fields := map[string]interface{}{
+		"usage_idle": float64(99),
+		"usage_busy": float64(1),
+	}
+	m, err := New("cpu", tags, fields, now, telegraf.Gauge)
+	assert.NoError(t, err)
+
+	assert.Equal(t, telegraf.Gauge, m.Type())
+	assert.Equal(t, tags, m.Tags())
+	assert.Equal(t, fields, m.Fields())
+	assert.Equal(t, "cpu", m.Name())
+	assert.Equal(t, now, m.Time())
+	assert.Equal(t, now.UnixNano(), m.UnixNano())
+}
+
+func TestNewCounterMetric(t *testing.T) {
+	now := time.Now()
+
+	tags := map[string]string{
+		"host":       "localhost",
+		"datacenter": "us-east-1",
+	}
+	fields := map[string]interface{}{
+		"usage_idle": float64(99),
+		"usage_busy": float64(1),
+	}
+	m, err := New("cpu", tags, fields, now, telegraf.Counter)
+	assert.NoError(t, err)
+
+	assert.Equal(t, telegraf.Counter, m.Type())
+	assert.Equal(t, tags, m.Tags())
+	assert.Equal(t, fields, m.Fields())
+	assert.Equal(t, "cpu", m.Name())
+	assert.Equal(t, now, m.Time())
+	assert.Equal(t, now.UnixNano(), m.UnixNano())
+}
+
+// test splitting metric into various max lengths
+func TestSplitMetric(t *testing.T) {
+	now := time.Unix(0, 1480940990034083306)
+	tags := map[string]string{
+		"host": "localhost",
+	}
+	fields := map[string]interface{}{
+		"float":  float64(100001),
+		"int":    int64(100001),
+		"bool":   true,
+		"false":  false,
+		"string": "test",
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	split80 := m.Split(80)
+	assert.Len(t, split80, 2)
+
+	split70 := m.Split(70)
+	assert.Len(t, split70, 3)
+
+	split60 := m.Split(60)
+	assert.Len(t, split60, 4)
+}
+
+// test splitting metric into various max lengths
+// use a simple regex check to verify that the split metrics are valid
+func TestSplitMetric_RegexVerify(t *testing.T) {
+	now := time.Unix(0, 1480940990034083306)
+	tags := map[string]string{
+		"host": "localhost",
+	}
+	fields := map[string]interface{}{
+		"foo":     float64(98934259085),
+		"bar":     float64(19385292),
+		"number":  float64(19385292),
+		"another": float64(19385292),
+		"n":       float64(19385292),
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	// verification regex
+	re := regexp.MustCompile(`cpu,host=localhost \w+=\d+(,\w+=\d+)* 1480940990034083306`)
+
+	split90 := m.Split(90)
+	assert.Len(t, split90, 2)
+	for _, splitM := range split90 {
+		assert.True(t, re.Match(splitM.Serialize()), splitM.String())
+	}
+
+	split70 := m.Split(70)
+	assert.Len(t, split70, 3)
+	for _, splitM := range split70 {
+		assert.True(t, re.Match(splitM.Serialize()), splitM.String())
+	}
+
+	split20 := m.Split(20)
+	assert.Len(t, split20, 5)
+	for _, splitM := range split20 {
+		assert.True(t, re.Match(splitM.Serialize()), splitM.String())
+	}
+}
+
+// test splitting metric even when given length is shorter than
+// shortest possible length
+// Split should split metric as short as possible, ie, 1 field per metric
+func TestSplitMetric_TooShort(t *testing.T) {
+	now := time.Unix(0, 1480940990034083306)
+	tags := map[string]string{
+		"host": "localhost",
+	}
+	fields := map[string]interface{}{
+		"float":  float64(100001),
+		"int":    int64(100001),
+		"bool":   true,
+		"false":  false,
+		"string": "test",
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	split := m.Split(10)
+	assert.Len(t, split, 5)
+	strings := make([]string, 5)
+	for i, splitM := range split {
+		strings[i] = splitM.String()
+	}
+
+	assert.Contains(t, strings, "cpu,host=localhost float=100001 1480940990034083306\n")
+	assert.Contains(t, strings, "cpu,host=localhost int=100001i 1480940990034083306\n")
+	assert.Contains(t, strings, "cpu,host=localhost bool=true 1480940990034083306\n")
+	assert.Contains(t, strings, "cpu,host=localhost false=false 1480940990034083306\n")
+	assert.Contains(t, strings, "cpu,host=localhost string=\"test\" 1480940990034083306\n")
+}
+
+func TestSplitMetric_NoOp(t *testing.T) {
+	now := time.Unix(0, 1480940990034083306)
+	tags := map[string]string{
+		"host": "localhost",
+	}
+	fields := map[string]interface{}{
+		"float":  float64(100001),
+		"int":    int64(100001),
+		"bool":   true,
+		"false":  false,
+		"string": "test",
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	split := m.Split(1000)
+	assert.Len(t, split, 1)
+	assert.Equal(t, m, split[0])
+}
+
+func TestSplitMetric_OneField(t *testing.T) {
+	now := time.Unix(0, 1480940990034083306)
+	tags := map[string]string{
+		"host": "localhost",
+	}
+	fields := map[string]interface{}{
+		"float": float64(100001),
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	assert.Equal(t, "cpu,host=localhost float=100001 1480940990034083306\n", m.String())
+
+	split := m.Split(1000)
+	assert.Len(t, split, 1)
+	assert.Equal(t, "cpu,host=localhost float=100001 1480940990034083306\n", split[0].String())
+
+	split = m.Split(1)
+	assert.Len(t, split, 1)
+	assert.Equal(t, "cpu,host=localhost float=100001 1480940990034083306\n", split[0].String())
+
+	split = m.Split(40)
+	assert.Len(t, split, 1)
+	assert.Equal(t, "cpu,host=localhost float=100001 1480940990034083306\n", split[0].String())
+}
+
+func TestNewMetricAggregate(t *testing.T) {
+	now := time.Now()
+
+	tags := map[string]string{
+		"host": "localhost",
+	}
+	fields := map[string]interface{}{
+		"usage_idle": float64(99),
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	assert.False(t, m.IsAggregate())
+	m.SetAggregate(true)
+	assert.True(t, m.IsAggregate())
+}
+
+func TestNewMetricString(t *testing.T) {
+	now := time.Now()
+
+	tags := map[string]string{
+		"host": "localhost",
+	}
+	fields := map[string]interface{}{
+		"usage_idle": float64(99),
+	}
+	m, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+
+	lineProto := fmt.Sprintf("cpu,host=localhost usage_idle=99 %d\n",
+		now.UnixNano())
+	assert.Equal(t, lineProto, m.String())
+}
+
+func TestNewMetricFailNaN(t *testing.T) {
+	now := time.Now()
+
+	tags := map[string]string{
+		"host": "localhost",
+	}
+	fields := map[string]interface{}{
+		"usage_idle": math.NaN(),
+	}
+
+	_, err := New("cpu", tags, fields, now)
+	assert.NoError(t, err)
+}
+
+func TestEmptyTagValueOrKey(t *testing.T) {
+	now := time.Now()
+
+	tags := map[string]string{
+		"host":     "localhost",
+		"emptytag": "",
+		"":         "valuewithoutkey",
+	}
+	fields := map[string]interface{}{
+		"usage_idle": float64(99),
+	}
+	m, err := New("cpu", tags, fields, now)
+
+	assert.True(t, m.HasTag("host"))
+	assert.False(t, m.HasTag("emptytag"))
+	assert.Equal(t,
+		fmt.Sprintf("cpu,host=localhost usage_idle=99 %d\n", now.UnixNano()),
+		m.String())
+
+	assert.NoError(t, err)
+
+}

metric/parse.go

@@ -0,0 +1,674 @@
+package metric
+
+import (
+	"bytes"
+	"errors"
+	"fmt"
+	"strconv"
+	"time"
+
+	"github.com/influxdata/telegraf"
+)
+
+var (
+	ErrInvalidNumber = errors.New("invalid number")
+)
+
+const (
+	// the number of characters for the largest possible int64 (9223372036854775807)
+	maxInt64Digits = 19
+
+	// the number of characters for the smallest possible int64 (-9223372036854775808)
+	minInt64Digits = 20
+
+	// the number of characters required for the largest float64 before a range check
+	// would occur during parsing
+	maxFloat64Digits = 25
+
+	// the number of characters required for smallest float64 before a range check occur
+	// would occur during parsing
+	minFloat64Digits = 27
+
+	MaxKeyLength = 65535
+)
+
+// The following constants allow us to specify which state to move to
+// next, when scanning sections of a Point.
+const (
+	tagKeyState = iota
+	tagValueState
+	fieldsState
+)
+
+func Parse(buf []byte) ([]telegraf.Metric, error) {
+	return ParseWithDefaultTimePrecision(buf, time.Now(), "")
+}
+
+func ParseWithDefaultTime(buf []byte, t time.Time) ([]telegraf.Metric, error) {
+	return ParseWithDefaultTimePrecision(buf, t, "")
+}
+
+func ParseWithDefaultTimePrecision(
+	buf []byte,
+	t time.Time,
+	precision string,
+) ([]telegraf.Metric, error) {
+	if len(buf) == 0 {
+		return []telegraf.Metric{}, nil
+	}
+	if len(buf) <= 6 {
+		return []telegraf.Metric{}, makeError("buffer too short", buf, 0)
+	}
+	metrics := make([]telegraf.Metric, 0, bytes.Count(buf, []byte("\n"))+1)
+	var errStr string
+	i := 0
+	for {
+		j := bytes.IndexByte(buf[i:], '\n')
+		if j == -1 {
+			break
+		}
+		if len(buf[i:i+j]) < 2 {
+			i += j + 1 // increment i past the previous newline
+			continue
+		}
+
+		m, err := parseMetric(buf[i:i+j], t, precision)
+		if err != nil {
+			i += j + 1 // increment i past the previous newline
+			errStr += " " + err.Error()
+			continue
+		}
+		i += j + 1 // increment i past the previous newline
+
+		metrics = append(metrics, m)
+	}
+
+	if len(errStr) > 0 {
+		return metrics, fmt.Errorf(errStr)
+	}
+	return metrics, nil
+}
+
+func parseMetric(buf []byte,
+	defaultTime time.Time,
+	precision string,
+) (telegraf.Metric, error) {
+	var dTime string
+	// scan the first block which is measurement[,tag1=value1,tag2=value=2...]
+	pos, key, err := scanKey(buf, 0)
+	if err != nil {
+		return nil, err
+	}
+
+	// measurement name is required
+	if len(key) == 0 {
+		return nil, fmt.Errorf("missing measurement")
+	}
+
+	if len(key) > MaxKeyLength {
+		return nil, fmt.Errorf("max key length exceeded: %v > %v", len(key), MaxKeyLength)
+	}
+
+	// scan the second block is which is field1=value1[,field2=value2,...]
+	pos, fields, err := scanFields(buf, pos)
+	if err != nil {
+		return nil, err
+	}
+
+	// at least one field is required
+	if len(fields) == 0 {
+		return nil, fmt.Errorf("missing fields")
+	}
+
+	// scan the last block which is an optional integer timestamp
+	pos, ts, err := scanTime(buf, pos)
+	if err != nil {
+		return nil, err
+	}
+
+	// apply precision multiplier
+	var nsec int64
+	multiplier := getPrecisionMultiplier(precision)
+	if multiplier > 1 {
+		tsint, err := parseIntBytes(ts, 10, 64)
+		if err != nil {
+			return nil, err
+		}
+
+		nsec := multiplier * tsint
+		ts = []byte(strconv.FormatInt(nsec, 10))
+	}
+
+	m := &metric{
+		fields: fields,
+		t:      ts,
+		nsec:   nsec,
+	}
+
+	// parse out the measurement name
+	// namei is the index at which the "name" ends
+	namei := indexUnescapedByte(key, ',')
+	if namei < 1 {
+		// no tags
+		m.name = key
+	} else {
+		m.name = key[0:namei]
+		m.tags = key[namei:]
+	}
+
+	if len(m.t) == 0 {
+		if len(dTime) == 0 {
+			dTime = fmt.Sprint(defaultTime.UnixNano())
+		}
+		// use default time
+		m.t = []byte(dTime)
+	}
+
+	// here we copy on return because this allows us to later call
+	// AddTag, AddField, RemoveTag, RemoveField, etc. without worrying about
+	// modifying 'tag' bytes having an affect on 'field' bytes, for example.
+	return m.Copy(), nil
+}
+
+// scanKey scans buf starting at i for the measurement and tag portion of the point.
+// It returns the ending position and the byte slice of key within buf.  If there
+// are tags, they will be sorted if they are not already.
+func scanKey(buf []byte, i int) (int, []byte, error) {
+	start := skipWhitespace(buf, i)
+	i = start
+
+	// First scan the Point's measurement.
+	state, i, err := scanMeasurement(buf, i)
+	if err != nil {
+		return i, buf[start:i], err
+	}
+
+	// Optionally scan tags if needed.
+	if state == tagKeyState {
+		i, err = scanTags(buf, i)
+		if err != nil {
+			return i, buf[start:i], err
+		}
+	}
+
+	return i, buf[start:i], nil
+}
+
+// scanMeasurement examines the measurement part of a Point, returning
+// the next state to move to, and the current location in the buffer.
+func scanMeasurement(buf []byte, i int) (int, int, error) {
+	// Check first byte of measurement, anything except a comma is fine.
+	// It can't be a space, since whitespace is stripped prior to this
+	// function call.
+	if i >= len(buf) || buf[i] == ',' {
+		return -1, i, makeError("missing measurement", buf, i)
+	}
+
+	for {
+		i++
+		if i >= len(buf) {
+			// cpu
+			return -1, i, makeError("missing fields", buf, i)
+		}
+
+		if buf[i-1] == '\\' {
+			// Skip character (it's escaped).
+			continue
+		}
+
+		// Unescaped comma; move onto scanning the tags.
+		if buf[i] == ',' {
+			return tagKeyState, i + 1, nil
+		}
+
+		// Unescaped space; move onto scanning the fields.
+		if buf[i] == ' ' {
+			// cpu value=1.0
+			return fieldsState, i, nil
+		}
+	}
+}
+
+// scanTags examines all the tags in a Point, keeping track of and
+// returning the updated indices slice, number of commas and location
+// in buf where to start examining the Point fields.
+func scanTags(buf []byte, i int) (int, error) {
+	var (
+		err   error
+		state = tagKeyState
+	)
+
+	for {
+		switch state {
+		case tagKeyState:
+			i, err = scanTagsKey(buf, i)
+			state = tagValueState // tag value always follows a tag key
+		case tagValueState:
+			state, i, err = scanTagsValue(buf, i)
+		case fieldsState:
+			return i, nil
+		}
+
+		if err != nil {
+			return i, err
+		}
+	}
+}
+
+// scanTagsKey scans each character in a tag key.
+func scanTagsKey(buf []byte, i int) (int, error) {
+	// First character of the key.
+	if i >= len(buf) || buf[i] == ' ' || buf[i] == ',' || buf[i] == '=' {
+		// cpu,{'', ' ', ',', '='}
+		return i, makeError("missing tag key", buf, i)
+	}
+
+	// Examine each character in the tag key until we hit an unescaped
+	// equals (the tag value), or we hit an error (i.e., unescaped
+	// space or comma).
+	for {
+		i++
+
+		// Either we reached the end of the buffer or we hit an
+		// unescaped comma or space.
+		if i >= len(buf) ||
+			((buf[i] == ' ' || buf[i] == ',') && buf[i-1] != '\\') {
+			// cpu,tag{'', ' ', ','}
+			return i, makeError("missing tag value", buf, i)
+		}
+
+		if buf[i] == '=' && buf[i-1] != '\\' {
+			// cpu,tag=
+			return i + 1, nil
+		}
+	}
+}
+
+// scanTagsValue scans each character in a tag value.
+func scanTagsValue(buf []byte, i int) (int, int, error) {
+	// Tag value cannot be empty.
+	if i >= len(buf) || buf[i] == ',' || buf[i] == ' ' {
+		// cpu,tag={',', ' '}
+		return -1, i, makeError("missing tag value", buf, i)
+	}
+
+	// Examine each character in the tag value until we hit an unescaped
+	// comma (move onto next tag key), an unescaped space (move onto
+	// fields), or we error out.
+	for {
+		i++
+		if i >= len(buf) {
+			// cpu,tag=value
+			return -1, i, makeError("missing fields", buf, i)
+		}
+
+		// An unescaped equals sign is an invalid tag value.
+		if buf[i] == '=' && buf[i-1] != '\\' {
+			// cpu,tag={'=', 'fo=o'}
+			return -1, i, makeError("invalid tag format", buf, i)
+		}
+
+		if buf[i] == ',' && buf[i-1] != '\\' {
+			// cpu,tag=foo,
+			return tagKeyState, i + 1, nil
+		}
+
+		// cpu,tag=foo value=1.0
+		// cpu, tag=foo\= value=1.0
+		if buf[i] == ' ' && buf[i-1] != '\\' {
+			return fieldsState, i, nil
+		}
+	}
+}
+
+// scanFields scans buf, starting at i for the fields section of a point.  It returns
+// the ending position and the byte slice of the fields within buf
+func scanFields(buf []byte, i int) (int, []byte, error) {
+	start := skipWhitespace(buf, i)
+	i = start
+	quoted := false
+
+	// tracks how many '=' we've seen
+	equals := 0
+
+	// tracks how many commas we've seen
+	commas := 0
+
+	for {
+		// reached the end of buf?
+		if i >= len(buf) {
+			break
+		}
+
+		// escaped characters?
+		if buf[i] == '\\' && i+1 < len(buf) {
+			i += 2
+			continue
+		}
+
+		// If the value is quoted, scan until we get to the end quote
+		// Only quote values in the field value since quotes are not significant
+		// in the field key
+		if buf[i] == '"' && equals > commas {
+			quoted = !quoted
+			i++
+			continue
+		}
+
+		// If we see an =, ensure that there is at least on char before and after it
+		if buf[i] == '=' && !quoted {
+			equals++
+
+			// check for "... =123" but allow "a\ =123"
+			if buf[i-1] == ' ' && buf[i-2] != '\\' {
+				return i, buf[start:i], makeError("missing field key", buf, i)
+			}
+
+			// check for "...a=123,=456" but allow "a=123,a\,=456"
+			if buf[i-1] == ',' && buf[i-2] != '\\' {
+				return i, buf[start:i], makeError("missing field key", buf, i)
+			}
+
+			// check for "... value="
+			if i+1 >= len(buf) {
+				return i, buf[start:i], makeError("missing field value", buf, i)
+			}
+
+			// check for "... value=,value2=..."
+			if buf[i+1] == ',' || buf[i+1] == ' ' {
+				return i, buf[start:i], makeError("missing field value", buf, i)
+			}
+
+			if isNumeric(buf[i+1]) || buf[i+1] == '-' || buf[i+1] == 'N' || buf[i+1] == 'n' {
+				var err error
+				i, err = scanNumber(buf, i+1)
+				if err != nil {
+					return i, buf[start:i], err
+				}
+				continue
+			}
+			// If next byte is not a double-quote, the value must be a boolean
+			if buf[i+1] != '"' {
+				var err error
+				i, _, err = scanBoolean(buf, i+1)
+				if err != nil {
+					return i, buf[start:i], err
+				}
+				continue
+			}
+		}
+
+		if buf[i] == ',' && !quoted {
+			commas++
+		}
+
+		// reached end of block?
+		if buf[i] == ' ' && !quoted {
+			break
+		}
+		i++
+	}
+
+	if quoted {
+		return i, buf[start:i], makeError("unbalanced quotes", buf, i)
+	}
+
+	// check that all field sections had key and values (e.g. prevent "a=1,b"
+	if equals == 0 || commas != equals-1 {
+		return i, buf[start:i], makeError("invalid field format", buf, i)
+	}
+
+	return i, buf[start:i], nil
+}
+
+// scanTime scans buf, starting at i for the time section of a point. It
+// returns the ending position and the byte slice of the timestamp within buf
+// and and error if the timestamp is not in the correct numeric format.
+func scanTime(buf []byte, i int) (int, []byte, error) {
+	start := skipWhitespace(buf, i)
+	i = start
+
+	for {
+		// reached the end of buf?
+		if i >= len(buf) {
+			break
+		}
+
+		// Reached end of block or trailing whitespace?
+		if buf[i] == '\n' || buf[i] == ' ' {
+			break
+		}
+
+		// Handle negative timestamps
+		if i == start && buf[i] == '-' {
+			i++
+			continue
+		}
+
+		// Timestamps should be integers, make sure they are so we don't need
+		// to actually  parse the timestamp until needed.
+		if buf[i] < '0' || buf[i] > '9' {
+			return i, buf[start:i], makeError("invalid timestamp", buf, i)
+		}
+		i++
+	}
+	return i, buf[start:i], nil
+}
+
+func isNumeric(b byte) bool {
+	return (b >= '0' && b <= '9') || b == '.'
+}
+
+// scanNumber returns the end position within buf, start at i after
+// scanning over buf for an integer, or float.  It returns an
+// error if a invalid number is scanned.
+func scanNumber(buf []byte, i int) (int, error) {
+	start := i
+	var isInt bool
+
+	// Is negative number?
+	if i < len(buf) && buf[i] == '-' {
+		i++
+		// There must be more characters now, as just '-' is illegal.
+		if i == len(buf) {
+			return i, ErrInvalidNumber
+		}
+	}
+
+	// how many decimal points we've see
+	decimal := false
+
+	// indicates the number is float in scientific notation
+	scientific := false
+
+	for {
+		if i >= len(buf) {
+			break
+		}
+
+		if buf[i] == ',' || buf[i] == ' ' {
+			break
+		}
+
+		if buf[i] == 'i' && i > start && !isInt {
+			isInt = true
+			i++
+			continue
+		}
+
+		if buf[i] == '.' {
+			// Can't have more than 1 decimal (e.g. 1.1.1 should fail)
+			if decimal {
+				return i, ErrInvalidNumber
+			}
+			decimal = true
+		}
+
+		// `e` is valid for floats but not as the first char
+		if i > start && (buf[i] == 'e' || buf[i] == 'E') {
+			scientific = true
+			i++
+			continue
+		}
+
+		// + and - are only valid at this point if they follow an e (scientific notation)
+		if (buf[i] == '+' || buf[i] == '-') && (buf[i-1] == 'e' || buf[i-1] == 'E') {
+			i++
+			continue
+		}
+
+		// NaN is an unsupported value
+		if i+2 < len(buf) && (buf[i] == 'N' || buf[i] == 'n') {
+			return i, ErrInvalidNumber
+		}
+
+		if !isNumeric(buf[i]) {
+			return i, ErrInvalidNumber
+		}
+		i++
+	}
+
+	if isInt && (decimal || scientific) {
+		return i, ErrInvalidNumber
+	}
+
+	numericDigits := i - start
+	if isInt {
+		numericDigits--
+	}
+	if decimal {
+		numericDigits--
+	}
+	if buf[start] == '-' {
+		numericDigits--
+	}
+
+	if numericDigits == 0 {
+		return i, ErrInvalidNumber
+	}
+
+	// It's more common that numbers will be within min/max range for their type but we need to prevent
+	// out or range numbers from being parsed successfully.  This uses some simple heuristics to decide
+	// if we should parse the number to the actual type.  It does not do it all the time because it incurs
+	// extra allocations and we end up converting the type again when writing points to disk.
+	if isInt {
+		// Make sure the last char is an 'i' for integers (e.g. 9i10 is not valid)
+		if buf[i-1] != 'i' {
+			return i, ErrInvalidNumber
+		}
+		// Parse the int to check bounds the number of digits could be larger than the max range
+		// We subtract 1 from the index to remove the `i` from our tests
+		if len(buf[start:i-1]) >= maxInt64Digits || len(buf[start:i-1]) >= minInt64Digits {
+			if _, err := parseIntBytes(buf[start:i-1], 10, 64); err != nil {
+				return i, makeError(fmt.Sprintf("unable to parse integer %s: %s", buf[start:i-1], err), buf, i)
+			}
+		}
+	} else {
+		// Parse the float to check bounds if it's scientific or the number of digits could be larger than the max range
+		if scientific || len(buf[start:i]) >= maxFloat64Digits || len(buf[start:i]) >= minFloat64Digits {
+			if _, err := parseFloatBytes(buf[start:i], 10); err != nil {
+				return i, makeError("invalid float", buf, i)
+			}
+		}
+	}
+
+	return i, nil
+}
+
+// scanBoolean returns the end position within buf, start at i after
+// scanning over buf for boolean. Valid values for a boolean are
+// t, T, true, TRUE, f, F, false, FALSE. It returns an error if a invalid boolean
+// is scanned.
+func scanBoolean(buf []byte, i int) (int, []byte, error) {
+	start := i
+
+	if i < len(buf) && (buf[i] != 't' && buf[i] != 'f' && buf[i] != 'T' && buf[i] != 'F') {
+		return i, buf[start:i], makeError("invalid value", buf, i)
+	}
+
+	i++
+	for {
+		if i >= len(buf) {
+			break
+		}
+
+		if buf[i] == ',' || buf[i] == ' ' {
+			break
+		}
+		i++
+	}
+
+	// Single char bool (t, T, f, F) is ok
+	if i-start == 1 {
+		return i, buf[start:i], nil
+	}
+
+	// length must be 4 for true or TRUE
+	if (buf[start] == 't' || buf[start] == 'T') && i-start != 4 {
+		return i, buf[start:i], makeError("invalid boolean", buf, i)
+	}
+
+	// length must be 5 for false or FALSE
+	if (buf[start] == 'f' || buf[start] == 'F') && i-start != 5 {
+		return i, buf[start:i], makeError("invalid boolean", buf, i)
+	}
+
+	// Otherwise
+	valid := false
+	switch buf[start] {
+	case 't':
+		valid = bytes.Equal(buf[start:i], []byte("true"))
+	case 'f':
+		valid = bytes.Equal(buf[start:i], []byte("false"))
+	case 'T':
+		valid = bytes.Equal(buf[start:i], []byte("TRUE")) || bytes.Equal(buf[start:i], []byte("True"))
+	case 'F':
+		valid = bytes.Equal(buf[start:i], []byte("FALSE")) || bytes.Equal(buf[start:i], []byte("False"))
+	}
+
+	if !valid {
+		return i, buf[start:i], makeError("invalid boolean", buf, i)
+	}
+
+	return i, buf[start:i], nil
+
+}
+
+// skipWhitespace returns the end position within buf, starting at i after
+// scanning over spaces in tags
+func skipWhitespace(buf []byte, i int) int {
+	for i < len(buf) {
+		if buf[i] != ' ' && buf[i] != '\t' && buf[i] != 0 {
+			break
+		}
+		i++
+	}
+	return i
+}
+
+// makeError is a helper function for making a metric parsing error.
+//   reason is the reason that the error occured.
+//   buf should be the current buffer we are parsing.
+//   i is the current index, to give some context on where in the buffer we are.
+func makeError(reason string, buf []byte, i int) error {
+	return fmt.Errorf("metric parsing error, reason: [%s], buffer: [%s], index: [%d]",
+		reason, buf, i)
+}
+
+// getPrecisionMultiplier will return a multiplier for the precision specified.
+func getPrecisionMultiplier(precision string) int64 {
+	d := time.Nanosecond
+	switch precision {
+	case "u":
+		d = time.Microsecond
+	case "ms":
+		d = time.Millisecond
+	case "s":
+		d = time.Second
+	case "m":
+		d = time.Minute
+	case "h":
+		d = time.Hour
+	}
+	return int64(d)
+}

metric/parse_test.go

@@ -0,0 +1,399 @@
+package metric
+
+import (
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+)
+
+const trues = `booltest b=T
+booltest b=t
+booltest b=True
+booltest b=TRUE
+booltest b=true
+`
+
+const falses = `booltest b=F
+booltest b=f
+booltest b=False
+booltest b=FALSE
+booltest b=false
+`
+
+const withEscapes = `w\,\ eather,host=local temp=99 1465839830100400200
+w\,eather,host=local temp=99 1465839830100400200
+weather,location=us\,midwest temperature=82 1465839830100400200
+weather,location=us-midwest temp\=rature=82 1465839830100400200
+weather,location\ place=us-midwest temperature=82 1465839830100400200
+weather,location=us-midwest temperature="too\"hot\"" 1465839830100400200
+`
+
+const withTimestamps = `cpu usage=99 1480595849000000000
+cpu usage=99 1480595850000000000
+cpu usage=99 1480595851700030000
+cpu usage=99 1480595852000000300
+`
+
+const sevenMetrics = `cpu,host=foo,datacenter=us-east idle=99,busy=1i,b=true,s="string"
+cpu,host=foo,datacenter=us-east idle=99,busy=1i,b=true,s="string"
+cpu,host=foo,datacenter=us-east idle=99,busy=1i,b=true,s="string"
+cpu,host=foo,datacenter=us-east idle=99,busy=1i,b=true,s="string"
+cpu,host=foo,datacenter=us-east idle=99,busy=1i,b=true,s="string"
+cpu,host=foo,datacenter=us-east idle=99,busy=1i,b=true,s="string"
+cpu,host=foo,datacenter=us-east idle=99,busy=1i,b=true,s="string"
+`
+
+const negMetrics = `weather,host=local temp=-99i,temp_float=-99.4 1465839830100400200
+`
+
+// some metrics are invalid
+const someInvalid = `cpu,host=foo,datacenter=us-east usage_idle=99,usage_busy=1
+cpu,host=foo,datacenter=us-east usage_idle=99,usage_busy=1
+cpu,host=foo,datacenter=us-east usage_idle=99,usage_busy=1
+cpu,cpu=cpu3, host=foo,datacenter=us-east usage_idle=99,usage_busy=1
+cpu,cpu=cpu4 , usage_idle=99,usage_busy=1
+cpu 1480595852000000300
+cpu usage=99 1480595852foobar300
+cpu,host=foo,datacenter=us-east usage_idle=99,usage_busy=1
+`
+
+func TestParse(t *testing.T) {
+	start := time.Now()
+	metrics, err := Parse([]byte(sevenMetrics))
+	assert.NoError(t, err)
+	assert.Len(t, metrics, 7)
+
+	// all metrics parsed together w/o a timestamp should have the same time.
+	firstTime := metrics[0].Time()
+	for _, m := range metrics {
+		assert.Equal(t,
+			map[string]interface{}{
+				"idle": float64(99),
+				"busy": int64(1),
+				"b":    true,
+				"s":    "string",
+			},
+			m.Fields(),
+		)
+		assert.Equal(t,
+			map[string]string{
+				"host":       "foo",
+				"datacenter": "us-east",
+			},
+			m.Tags(),
+		)
+		assert.True(t, m.Time().After(start))
+		assert.True(t, m.Time().Equal(firstTime))
+	}
+}
+
+func TestParseNegNumbers(t *testing.T) {
+	metrics, err := Parse([]byte(negMetrics))
+	assert.NoError(t, err)
+	assert.Len(t, metrics, 1)
+
+	assert.Equal(t,
+		map[string]interface{}{
+			"temp":       int64(-99),
+			"temp_float": float64(-99.4),
+		},
+		metrics[0].Fields(),
+	)
+	assert.Equal(t,
+		map[string]string{
+			"host": "local",
+		},
+		metrics[0].Tags(),
+	)
+}
+
+func TestParseErrors(t *testing.T) {
+	start := time.Now()
+	metrics, err := Parse([]byte(someInvalid))
+	assert.Error(t, err)
+	assert.Len(t, metrics, 4)
+
+	// all metrics parsed together w/o a timestamp should have the same time.
+	firstTime := metrics[0].Time()
+	for _, m := range metrics {
+		assert.Equal(t,
+			map[string]interface{}{
+				"usage_idle": float64(99),
+				"usage_busy": float64(1),
+			},
+			m.Fields(),
+		)
+		assert.Equal(t,
+			map[string]string{
+				"host":       "foo",
+				"datacenter": "us-east",
+			},
+			m.Tags(),
+		)
+		assert.True(t, m.Time().After(start))
+		assert.True(t, m.Time().Equal(firstTime))
+	}
+}
+
+func TestParseWithTimestamps(t *testing.T) {
+	metrics, err := Parse([]byte(withTimestamps))
+	assert.NoError(t, err)
+	assert.Len(t, metrics, 4)
+
+	expectedTimestamps := []time.Time{
+		time.Unix(0, 1480595849000000000),
+		time.Unix(0, 1480595850000000000),
+		time.Unix(0, 1480595851700030000),
+		time.Unix(0, 1480595852000000300),
+	}
+
+	// all metrics parsed together w/o a timestamp should have the same time.
+	for i, m := range metrics {
+		assert.Equal(t,
+			map[string]interface{}{
+				"usage": float64(99),
+			},
+			m.Fields(),
+		)
+		assert.True(t, m.Time().Equal(expectedTimestamps[i]))
+	}
+}
+
+func TestParseEscapes(t *testing.T) {
+	metrics, err := Parse([]byte(withEscapes))
+	assert.NoError(t, err)
+	assert.Len(t, metrics, 6)
+
+	tests := []struct {
+		name   string
+		fields map[string]interface{}
+		tags   map[string]string
+	}{
+		{
+			name:   `w, eather`,
+			fields: map[string]interface{}{"temp": float64(99)},
+			tags:   map[string]string{"host": "local"},
+		},
+		{
+			name:   `w,eather`,
+			fields: map[string]interface{}{"temp": float64(99)},
+			tags:   map[string]string{"host": "local"},
+		},
+		{
+			name:   `weather`,
+			fields: map[string]interface{}{"temperature": float64(82)},
+			tags:   map[string]string{"location": `us,midwest`},
+		},
+		{
+			name:   `weather`,
+			fields: map[string]interface{}{`temp=rature`: float64(82)},
+			tags:   map[string]string{"location": `us-midwest`},
+		},
+		{
+			name:   `weather`,
+			fields: map[string]interface{}{"temperature": float64(82)},
+			tags:   map[string]string{`location place`: `us-midwest`},
+		},
+		{
+			name:   `weather`,
+			fields: map[string]interface{}{`temperature`: `too"hot"`},
+			tags:   map[string]string{"location": `us-midwest`},
+		},
+	}
+
+	for i, test := range tests {
+		assert.Equal(t, test.name, metrics[i].Name())
+		assert.Equal(t, test.fields, metrics[i].Fields())
+		assert.Equal(t, test.tags, metrics[i].Tags())
+	}
+}
+
+func TestParseTrueBooleans(t *testing.T) {
+	metrics, err := Parse([]byte(trues))
+	assert.NoError(t, err)
+	assert.Len(t, metrics, 5)
+
+	for _, metric := range metrics {
+		assert.Equal(t, "booltest", metric.Name())
+		assert.Equal(t, true, metric.Fields()["b"])
+	}
+}
+
+func TestParseFalseBooleans(t *testing.T) {
+	metrics, err := Parse([]byte(falses))
+	assert.NoError(t, err)
+	assert.Len(t, metrics, 5)
+
+	for _, metric := range metrics {
+		assert.Equal(t, "booltest", metric.Name())
+		assert.Equal(t, false, metric.Fields()["b"])
+	}
+}
+
+func TestParsePointBadNumber(t *testing.T) {
+	for _, tt := range []string{
+		"cpu v=- ",
+		"cpu v=-i ",
+		"cpu v=-. ",
+		"cpu v=. ",
+		"cpu v=1.0i ",
+		"cpu v=1ii ",
+		"cpu v=1a ",
+		"cpu v=-e-e-e ",
+		"cpu v=42+3 ",
+		"cpu v= ",
+	} {
+		_, err := Parse([]byte(tt + "\n"))
+		assert.Error(t, err, tt)
+	}
+}
+
+func TestParseTagsMissingParts(t *testing.T) {
+	for _, tt := range []string{
+		`cpu,host`,
+		`cpu,host,`,
+		`cpu,host=`,
+		`cpu,f=oo=bar value=1`,
+		`cpu,host value=1i`,
+		`cpu,host=serverA,region value=1i`,
+		`cpu,host=serverA,region= value=1i`,
+		`cpu,host=serverA,region=,zone=us-west value=1i`,
+		`cpu, value=1`,
+		`cpu, ,,`,
+		`cpu,,,`,
+		`cpu,host=serverA,=us-east value=1i`,
+		`cpu,host=serverAa\,,=us-east value=1i`,
+		`cpu,host=serverA\,,=us-east value=1i`,
+		`cpu, =serverA value=1i`,
+	} {
+		_, err := Parse([]byte(tt + "\n"))
+		assert.Error(t, err, tt)
+	}
+}
+
+func TestParsePointWhitespace(t *testing.T) {
+	for _, tt := range []string{
+		`cpu    value=1.0 1257894000000000000`,
+		`cpu value=1.0     1257894000000000000`,
+		`cpu      value=1.0     1257894000000000000`,
+		`cpu value=1.0 1257894000000000000   `,
+	} {
+		m, err := Parse([]byte(tt + "\n"))
+		assert.NoError(t, err, tt)
+		assert.Equal(t, "cpu", m[0].Name())
+		assert.Equal(t, map[string]interface{}{"value": float64(1)}, m[0].Fields())
+	}
+}
+
+func TestParsePointInvalidFields(t *testing.T) {
+	for _, tt := range []string{
+		"test,foo=bar a=101,=value",
+		"test,foo=bar =value",
+		"test,foo=bar a=101,key=",
+		"test,foo=bar key=",
+		`test,foo=bar a=101,b="foo`,
+	} {
+		_, err := Parse([]byte(tt + "\n"))
+		assert.Error(t, err, tt)
+	}
+}
+
+func TestParsePointNoFields(t *testing.T) {
+	for _, tt := range []string{
+		"cpu_load_short,host=server01,region=us-west",
+		"very_long_measurement_name",
+		"cpu,host==",
+		"============",
+		"cpu",
+		"cpu\n\n\n\n\n\n\n",
+		"                ",
+	} {
+		_, err := Parse([]byte(tt + "\n"))
+		assert.Error(t, err, tt)
+	}
+}
+
+// a b=1 << this is the shortest possible metric
+// any shorter is just ignored
+func TestParseBufTooShort(t *testing.T) {
+	for _, tt := range []string{
+		"",
+		"a",
+		"a ",
+		"a b=",
+	} {
+		_, err := Parse([]byte(tt + "\n"))
+		assert.Error(t, err, tt)
+	}
+}
+
+func TestParseInvalidBooleans(t *testing.T) {
+	for _, tt := range []string{
+		"test b=tru",
+		"test b=fals",
+		"test b=faLse",
+		"test q=foo",
+		"test b=lambchops",
+	} {
+		_, err := Parse([]byte(tt + "\n"))
+		assert.Error(t, err, tt)
+	}
+}
+
+func TestParseInvalidNumbers(t *testing.T) {
+	for _, tt := range []string{
+		"test b=-",
+		"test b=1.1.1",
+		"test b=nan",
+		"test b=9i10",
+		"test b=9999999999999999999i",
+	} {
+		_, err := Parse([]byte(tt + "\n"))
+		assert.Error(t, err, tt)
+	}
+}
+
+func TestParseNegativeTimestamps(t *testing.T) {
+	for _, tt := range []string{
+		"test foo=101 -1257894000000000000",
+	} {
+		metrics, err := Parse([]byte(tt + "\n"))
+		assert.NoError(t, err, tt)
+		assert.True(t, metrics[0].Time().Equal(time.Unix(0, -1257894000000000000)))
+	}
+}
+
+func TestParsePrecision(t *testing.T) {
+	for _, tt := range []struct {
+		line      string
+		precision string
+		expected  int64
+	}{
+		{"test v=42 1491847420", "s", 1491847420000000000},
+		{"test v=42 1491847420123", "ms", 1491847420123000000},
+		{"test v=42 1491847420123456", "u", 1491847420123456000},
+		{"test v=42 1491847420123456789", "ns", 1491847420123456789},
+
+		{"test v=42 1491847420123456789", "1s", 1491847420123456789},
+		{"test v=42 1491847420123456789", "asdf", 1491847420123456789},
+	} {
+		metrics, err := ParseWithDefaultTimePrecision(
+			[]byte(tt.line+"\n"), time.Now(), tt.precision)
+		assert.NoError(t, err, tt)
+		assert.Equal(t, tt.expected, metrics[0].UnixNano())
+	}
+}
+
+func TestParseMaxKeyLength(t *testing.T) {
+	key := ""
+	for {
+		if len(key) > MaxKeyLength {
+			break
+		}
+		key += "test"
+	}
+
+	_, err := Parse([]byte(key + " value=1\n"))
+	assert.Error(t, err)
+}

metric/reader.go

@@ -0,0 +1,155 @@
+package metric
+
+import (
+	"io"
+
+	"github.com/influxdata/telegraf"
+)
+
+type state int
+
+const (
+	_ state = iota
+	// normal state copies whole metrics into the given buffer until we can't
+	// fit the next metric.
+	normal
+	// split state means that we have a metric that we were able to split, so
+	// that we can fit it into multiple metrics (and calls to Read)
+	split
+	// overflow state means that we have a metric that didn't fit into a single
+	// buffer, and needs to be split across multiple calls to Read.
+	overflow
+	// splitOverflow state means that a split metric didn't fit into a single
+	// buffer, and needs to be split across multiple calls to Read.
+	splitOverflow
+	// done means we're done reading metrics, and now always return (0, io.EOF)
+	done
+)
+
+type reader struct {
+	metrics      []telegraf.Metric
+	splitMetrics []telegraf.Metric
+	buf          []byte
+	state        state
+
+	// metric index
+	iM int
+	// split metric index
+	iSM int
+	// buffer index
+	iB int
+}
+
+func NewReader(metrics []telegraf.Metric) io.Reader {
+	return &reader{
+		metrics: metrics,
+		state:   normal,
+	}
+}
+
+func (r *reader) Read(p []byte) (n int, err error) {
+	var i int
+	switch r.state {
+	case done:
+		return 0, io.EOF
+	case normal:
+		for {
+			// this for-loop is the sunny-day scenario, where we are given a
+			// buffer that is large enough to hold at least a single metric.
+			// all of the cases below it are edge-cases.
+			if r.metrics[r.iM].Len() < len(p[i:]) {
+				i += r.metrics[r.iM].SerializeTo(p[i:])
+			} else {
+				break
+			}
+			r.iM++
+			if r.iM == len(r.metrics) {
+				r.state = done
+				return i, io.EOF
+			}
+		}
+
+		// if we haven't written any bytes, check if we can split the current
+		// metric into multiple full metrics at a smaller size.
+		if i == 0 {
+			tmp := r.metrics[r.iM].Split(len(p))
+			if len(tmp) > 1 {
+				r.splitMetrics = tmp
+				r.state = split
+				if r.splitMetrics[0].Len() < len(p) {
+					i += r.splitMetrics[0].SerializeTo(p)
+					r.iSM = 1
+				} else {
+					// splitting didn't quite work, so we'll drop down and
+					// overflow the metric.
+					r.state = normal
+					r.iSM = 0
+				}
+			}
+		}
+
+		// if we haven't written any bytes and we're not at the end of the metrics
+		// slice, then it means we have a single metric that is larger than the
+		// provided buffer.
+		if i == 0 {
+			r.buf = r.metrics[r.iM].Serialize()
+			i += copy(p, r.buf[r.iB:])
+			r.iB += i
+			r.state = overflow
+		}
+
+	case split:
+		if r.splitMetrics[r.iSM].Len() < len(p) {
+			// write the current split metric
+			i += r.splitMetrics[r.iSM].SerializeTo(p)
+			r.iSM++
+			if r.iSM >= len(r.splitMetrics) {
+				// done writing the current split metrics
+				r.iSM = 0
+				r.iM++
+				if r.iM == len(r.metrics) {
+					r.state = done
+					return i, io.EOF
+				}
+				r.state = normal
+			}
+		} else {
+			// This would only happen if we split the metric, and then a
+			// subsequent buffer was smaller than the initial one given,
+			// so that our split metric no longer fits.
+			r.buf = r.splitMetrics[r.iSM].Serialize()
+			i += copy(p, r.buf[r.iB:])
+			r.iB += i
+			r.state = splitOverflow
+		}
+
+	case splitOverflow:
+		i = copy(p, r.buf[r.iB:])
+		r.iB += i
+		if r.iB >= len(r.buf) {
+			r.iB = 0
+			r.iSM++
+			if r.iSM == len(r.splitMetrics) {
+				r.iM++
+				r.state = normal
+			} else {
+				r.state = split
+			}
+		}
+
+	case overflow:
+		i = copy(p, r.buf[r.iB:])
+		r.iB += i
+		if r.iB >= len(r.buf) {
+			r.iB = 0
+			r.iM++
+			if r.iM == len(r.metrics) {
+				r.state = done
+				return i, io.EOF
+			}
+			r.state = normal
+		}
+	}
+
+	return i, nil
+}

metric/reader_test.go

@@ -0,0 +1,487 @@
+package metric
+
+import (
+	"io"
+	"io/ioutil"
+	"regexp"
+	"testing"
+	"time"
+
+	"github.com/influxdata/telegraf"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func BenchmarkMetricReader(b *testing.B) {
+	metrics := make([]telegraf.Metric, 10)
+	for i := 0; i < 10; i++ {
+		metrics[i], _ = New("foo", map[string]string{},
+			map[string]interface{}{"value": int64(1)}, time.Now())
+	}
+	for n := 0; n < b.N; n++ {
+		r := NewReader(metrics)
+		io.Copy(ioutil.Discard, r)
+	}
+}
+
+func TestMetricReader(t *testing.T) {
+	ts := time.Unix(1481032190, 0)
+	metrics := make([]telegraf.Metric, 10)
+	for i := 0; i < 10; i++ {
+		metrics[i], _ = New("foo", map[string]string{},
+			map[string]interface{}{"value": int64(1)}, ts)
+	}
+
+	r := NewReader(metrics)
+
+	buf := make([]byte, 35)
+	for i := 0; i < 10; i++ {
+		n, err := r.Read(buf)
+		if err != nil {
+			assert.True(t, err == io.EOF, err.Error())
+		}
+		assert.Equal(t, 33, n)
+		assert.Equal(t, "foo value=1i 1481032190000000000\n", string(buf[0:n]))
+	}
+
+	// reader should now be done, and always return 0, io.EOF
+	for i := 0; i < 10; i++ {
+		n, err := r.Read(buf)
+		assert.True(t, err == io.EOF, err.Error())
+		assert.Equal(t, 0, n)
+	}
+}
+
+func TestMetricReader_OverflowMetric(t *testing.T) {
+	ts := time.Unix(1481032190, 0)
+	m, _ := New("foo", map[string]string{},
+		map[string]interface{}{"value": int64(10)}, ts)
+	metrics := []telegraf.Metric{m}
+
+	r := NewReader(metrics)
+	buf := make([]byte, 5)
+
+	tests := []struct {
+		exp string
+		err error
+		n   int
+	}{
+		{
+			"foo v",
+			nil,
+			5,
+		},
+		{
+			"alue=",
+			nil,
+			5,
+		},
+		{
+			"10i 1",
+			nil,
+			5,
+		},
+		{
+			"48103",
+			nil,
+			5,
+		},
+		{
+			"21900",
+			nil,
+			5,
+		},
+		{
+			"00000",
+			nil,
+			5,
+		},
+		{
+			"000\n",
+			io.EOF,
+			4,
+		},
+		{
+			"",
+			io.EOF,
+			0,
+		},
+	}
+
+	for _, test := range tests {
+		n, err := r.Read(buf)
+		assert.Equal(t, test.n, n)
+		assert.Equal(t, test.exp, string(buf[0:n]))
+		assert.Equal(t, test.err, err)
+	}
+}
+
+func TestMetricReader_OverflowMultipleMetrics(t *testing.T) {
+	ts := time.Unix(1481032190, 0)
+	m, _ := New("foo", map[string]string{},
+		map[string]interface{}{"value": int64(10)}, ts)
+	metrics := []telegraf.Metric{m, m.Copy()}
+
+	r := NewReader(metrics)
+	buf := make([]byte, 10)
+
+	tests := []struct {
+		exp string
+		err error
+		n   int
+	}{
+		{
+			"foo value=",
+			nil,
+			10,
+		},
+		{
+			"10i 148103",
+			nil,
+			10,
+		},
+		{
+			"2190000000",
+			nil,
+			10,
+		},
+		{
+			"000\n",
+			nil,
+			4,
+		},
+		{
+			"foo value=",
+			nil,
+			10,
+		},
+		{
+			"10i 148103",
+			nil,
+			10,
+		},
+		{
+			"2190000000",
+			nil,
+			10,
+		},
+		{
+			"000\n",
+			io.EOF,
+			4,
+		},
+		{
+			"",
+			io.EOF,
+			0,
+		},
+	}
+
+	for _, test := range tests {
+		n, err := r.Read(buf)
+		assert.Equal(t, test.n, n)
+		assert.Equal(t, test.exp, string(buf[0:n]))
+		assert.Equal(t, test.err, err)
+	}
+}
+
+// test splitting a metric
+func TestMetricReader_SplitMetric(t *testing.T) {
+	ts := time.Unix(1481032190, 0)
+	m1, _ := New("foo", map[string]string{},
+		map[string]interface{}{
+			"value1": int64(10),
+			"value2": int64(10),
+			"value3": int64(10),
+			"value4": int64(10),
+			"value5": int64(10),
+			"value6": int64(10),
+		},
+		ts,
+	)
+	metrics := []telegraf.Metric{m1}
+
+	r := NewReader(metrics)
+	buf := make([]byte, 60)
+
+	tests := []struct {
+		expRegex string
+		err      error
+		n        int
+	}{
+		{
+			`foo value\d=10i,value\d=10i,value\d=10i 1481032190000000000\n`,
+			nil,
+			57,
+		},
+		{
+			`foo value\d=10i,value\d=10i,value\d=10i 1481032190000000000\n`,
+			io.EOF,
+			57,
+		},
+		{
+			"",
+			io.EOF,
+			0,
+		},
+	}
+
+	for _, test := range tests {
+		n, err := r.Read(buf)
+		assert.Equal(t, test.n, n)
+		re := regexp.MustCompile(test.expRegex)
+		assert.True(t, re.MatchString(string(buf[0:n])), string(buf[0:n]))
+		assert.Equal(t, test.err, err)
+	}
+}
+
+// test an array with one split metric and one unsplit
+func TestMetricReader_SplitMetric2(t *testing.T) {
+	ts := time.Unix(1481032190, 0)
+	m1, _ := New("foo", map[string]string{},
+		map[string]interface{}{
+			"value1": int64(10),
+			"value2": int64(10),
+			"value3": int64(10),
+			"value4": int64(10),
+			"value5": int64(10),
+			"value6": int64(10),
+		},
+		ts,
+	)
+	m2, _ := New("foo", map[string]string{},
+		map[string]interface{}{
+			"value1": int64(10),
+		},
+		ts,
+	)
+	metrics := []telegraf.Metric{m1, m2}
+
+	r := NewReader(metrics)
+	buf := make([]byte, 60)
+
+	tests := []struct {
+		expRegex string
+		err      error
+		n        int
+	}{
+		{
+			`foo value\d=10i,value\d=10i,value\d=10i 1481032190000000000\n`,
+			nil,
+			57,
+		},
+		{
+			`foo value\d=10i,value\d=10i,value\d=10i 1481032190000000000\n`,
+			nil,
+			57,
+		},
+		{
+			`foo value1=10i 1481032190000000000\n`,
+			io.EOF,
+			35,
+		},
+		{
+			"",
+			io.EOF,
+			0,
+		},
+	}
+
+	for _, test := range tests {
+		n, err := r.Read(buf)
+		assert.Equal(t, test.n, n)
+		re := regexp.MustCompile(test.expRegex)
+		assert.True(t, re.MatchString(string(buf[0:n])), string(buf[0:n]))
+		assert.Equal(t, test.err, err)
+	}
+}
+
+// test split that results in metrics that are still too long, which results in
+// the reader falling back to regular overflow.
+func TestMetricReader_SplitMetricTooLong(t *testing.T) {
+	ts := time.Unix(1481032190, 0)
+	m1, _ := New("foo", map[string]string{},
+		map[string]interface{}{
+			"value1": int64(10),
+			"value2": int64(10),
+		},
+		ts,
+	)
+	metrics := []telegraf.Metric{m1}
+
+	r := NewReader(metrics)
+	buf := make([]byte, 30)
+
+	tests := []struct {
+		expRegex string
+		err      error
+		n        int
+	}{
+		{
+			`foo value\d=10i,value\d=10i 1481`,
+			nil,
+			30,
+		},
+		{
+			`032190000000000\n`,
+			io.EOF,
+			16,
+		},
+		{
+			"",
+			io.EOF,
+			0,
+		},
+	}
+
+	for _, test := range tests {
+		n, err := r.Read(buf)
+		assert.Equal(t, test.n, n)
+		re := regexp.MustCompile(test.expRegex)
+		assert.True(t, re.MatchString(string(buf[0:n])), string(buf[0:n]))
+		assert.Equal(t, test.err, err)
+	}
+}
+
+// test split with a changing buffer size in the middle of subsequent calls
+// to Read
+func TestMetricReader_SplitMetricChangingBuffer(t *testing.T) {
+	ts := time.Unix(1481032190, 0)
+	m1, _ := New("foo", map[string]string{},
+		map[string]interface{}{
+			"value1": int64(10),
+			"value2": int64(10),
+			"value3": int64(10),
+		},
+		ts,
+	)
+	m2, _ := New("foo", map[string]string{},
+		map[string]interface{}{
+			"value1": int64(10),
+		},
+		ts,
+	)
+	metrics := []telegraf.Metric{m1, m2}
+
+	r := NewReader(metrics)
+
+	tests := []struct {
+		expRegex string
+		err      error
+		n        int
+		buf      []byte
+	}{
+		{
+			`foo value\d=10i 1481032190000000000\n`,
+			nil,
+			35,
+			make([]byte, 36),
+		},
+		{
+			`foo value\d=10i 148103219000000`,
+			nil,
+			30,
+			make([]byte, 30),
+		},
+		{
+			`0000\n`,
+			nil,
+			5,
+			make([]byte, 30),
+		},
+		{
+			`foo value\d=10i 1481032190000000000\n`,
+			nil,
+			35,
+			make([]byte, 36),
+		},
+		{
+			`foo value1=10i 1481032190000000000\n`,
+			io.EOF,
+			35,
+			make([]byte, 36),
+		},
+		{
+			"",
+			io.EOF,
+			0,
+			make([]byte, 36),
+		},
+	}
+
+	for _, test := range tests {
+		n, err := r.Read(test.buf)
+		assert.Equal(t, test.n, n, test.expRegex)
+		re := regexp.MustCompile(test.expRegex)
+		assert.True(t, re.MatchString(string(test.buf[0:n])), string(test.buf[0:n]))
+		assert.Equal(t, test.err, err, test.expRegex)
+	}
+}
+
+// test split with a changing buffer size in the middle of subsequent calls
+// to Read
+func TestMetricReader_SplitMetricChangingBuffer2(t *testing.T) {
+	ts := time.Unix(1481032190, 0)
+	m1, _ := New("foo", map[string]string{},
+		map[string]interface{}{
+			"value1": int64(10),
+			"value2": int64(10),
+		},
+		ts,
+	)
+	m2, _ := New("foo", map[string]string{},
+		map[string]interface{}{
+			"value1": int64(10),
+		},
+		ts,
+	)
+	metrics := []telegraf.Metric{m1, m2}
+
+	r := NewReader(metrics)
+
+	tests := []struct {
+		expRegex string
+		err      error
+		n        int
+		buf      []byte
+	}{
+		{
+			`foo value\d=10i 1481032190000000000\n`,
+			nil,
+			35,
+			make([]byte, 36),
+		},
+		{
+			`foo value\d=10i 148103219000000`,
+			nil,
+			30,
+			make([]byte, 30),
+		},
+		{
+			`0000\n`,
+			nil,
+			5,
+			make([]byte, 30),
+		},
+		{
+			`foo value1=10i 1481032190000000000\n`,
+			io.EOF,
+			35,
+			make([]byte, 36),
+		},
+		{
+			"",
+			io.EOF,
+			0,
+			make([]byte, 36),
+		},
+	}
+
+	for _, test := range tests {
+		n, err := r.Read(test.buf)
+		assert.Equal(t, test.n, n, test.expRegex)
+		re := regexp.MustCompile(test.expRegex)
+		assert.True(t, re.MatchString(string(test.buf[0:n])), string(test.buf[0:n]))
+		assert.Equal(t, test.err, err, test.expRegex)
+	}
+}

metric_test.go

@@ -1,111 +0,0 @@
-package telegraf
-
-import (
-	"fmt"
-	"math"
-	"testing"
-	"time"
-
-	"github.com/stretchr/testify/assert"
-)
-
-func TestNewMetric(t *testing.T) {
-	now := time.Now()
-
-	tags := map[string]string{
-		"host":       "localhost",
-		"datacenter": "us-east-1",
-	}
-	fields := map[string]interface{}{
-		"usage_idle": float64(99),
-		"usage_busy": float64(1),
-	}
-	m, err := NewMetric("cpu", tags, fields, now)
-	assert.NoError(t, err)
-
-	assert.Equal(t, Untyped, m.Type())
-	assert.Equal(t, tags, m.Tags())
-	assert.Equal(t, fields, m.Fields())
-	assert.Equal(t, "cpu", m.Name())
-	assert.Equal(t, now, m.Time())
-	assert.Equal(t, now.UnixNano(), m.UnixNano())
-}
-
-func TestNewGaugeMetric(t *testing.T) {
-	now := time.Now()
-
-	tags := map[string]string{
-		"host":       "localhost",
-		"datacenter": "us-east-1",
-	}
-	fields := map[string]interface{}{
-		"usage_idle": float64(99),
-		"usage_busy": float64(1),
-	}
-	m, err := NewGaugeMetric("cpu", tags, fields, now)
-	assert.NoError(t, err)
-
-	assert.Equal(t, Gauge, m.Type())
-	assert.Equal(t, tags, m.Tags())
-	assert.Equal(t, fields, m.Fields())
-	assert.Equal(t, "cpu", m.Name())
-	assert.Equal(t, now, m.Time())
-	assert.Equal(t, now.UnixNano(), m.UnixNano())
-}
-
-func TestNewCounterMetric(t *testing.T) {
-	now := time.Now()
-
-	tags := map[string]string{
-		"host":       "localhost",
-		"datacenter": "us-east-1",
-	}
-	fields := map[string]interface{}{
-		"usage_idle": float64(99),
-		"usage_busy": float64(1),
-	}
-	m, err := NewCounterMetric("cpu", tags, fields, now)
-	assert.NoError(t, err)
-
-	assert.Equal(t, Counter, m.Type())
-	assert.Equal(t, tags, m.Tags())
-	assert.Equal(t, fields, m.Fields())
-	assert.Equal(t, "cpu", m.Name())
-	assert.Equal(t, now, m.Time())
-	assert.Equal(t, now.UnixNano(), m.UnixNano())
-}
-
-func TestNewMetricString(t *testing.T) {
-	now := time.Now()
-
-	tags := map[string]string{
-		"host": "localhost",
-	}
-	fields := map[string]interface{}{
-		"usage_idle": float64(99),
-	}
-	m, err := NewMetric("cpu", tags, fields, now)
-	assert.NoError(t, err)
-
-	lineProto := fmt.Sprintf("cpu,host=localhost usage_idle=99 %d",
-		now.UnixNano())
-	assert.Equal(t, lineProto, m.String())
-
-	lineProtoPrecision := fmt.Sprintf("cpu,host=localhost usage_idle=99 %d",
-		now.Unix())
-	assert.Equal(t, lineProtoPrecision, m.PrecisionString("s"))
-}
-
-func TestNewMetricFailNaN(t *testing.T) {
-	now := time.Now()
-
-	tags := map[string]string{
-		"host": "localhost",
-	}
-	fields := map[string]interface{}{
-		"usage_idle": math.NaN(),
-	}
-
-	_, err := NewMetric("cpu", tags, fields, now)
-	assert.Error(t, err)
-}

plugins/aggregators/minmax/README.md

@@ -0,0 +1,42 @@
+# MinMax Aggregator Plugin
+
+The minmax aggregator plugin aggregates min & max values of each field it sees,
+emitting the aggrate every `period` seconds.
+
+### Configuration:
+
+```toml
+# Keep the aggregate min/max of each metric passing through.
+[[aggregators.minmax]]
+  ## General Aggregator Arguments:
+  ## The period on which to flush & clear the aggregator.
+  period = "30s"
+  ## If true, the original metric will be dropped by the
+  ## aggregator and will not get sent to the output plugins.
+  drop_original = false
+```
+
+### Measurements & Fields:
+
+- measurement1
+    - field1_max
+    - field1_min
+
+### Tags:
+
+No tags are applied by this aggregator.
+
+### Example Output:
+
+```
+$ telegraf --config telegraf.conf --quiet
+system,host=tars load1=1.72 1475583980000000000
+system,host=tars load1=1.6 1475583990000000000
+system,host=tars load1=1.66 1475584000000000000
+system,host=tars load1=1.63 1475584010000000000
+system,host=tars load1_max=1.72,load1_min=1.6 1475584010000000000
+system,host=tars load1=1.46 1475584020000000000
+system,host=tars load1=1.39 1475584030000000000
+system,host=tars load1=1.41 1475584040000000000
+system,host=tars load1_max=1.46,load1_min=1.39 1475584040000000000
+```

plugins/aggregators/minmax/minmax_test.go

@@ -4,11 +4,11 @@ import (
 	"testing"
 	"time"
 
-	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/metric"
 	"github.com/influxdata/telegraf/testutil"
 )
 
-var m1, _ = telegraf.NewMetric("m1",
+var m1, _ = metric.New("m1",
 	map[string]string{"foo": "bar"},
 	map[string]interface{}{
 		"a": int64(1),
@@ -24,7 +24,7 @@ var m1, _ = telegraf.NewMetric("m1",
 	},
 	time.Now(),
 )
-var m2, _ = telegraf.NewMetric("m1",
+var m2, _ = metric.New("m1",
 	map[string]string{"foo": "bar"},
 	map[string]interface{}{
 		"a":        int64(1),

plugins/inputs/EXAMPLE_README.md

@@ -1,6 +1,8 @@
 # Example Input Plugin
 
-The example plugin gathers metrics about example things
+The example plugin gathers metrics about example things.  This description
+explains at a high level what the plugin does and provides links to where
+additional information can be found.
 
 ### Configuration:
 
@@ -12,7 +14,8 @@ The example plugin gathers metrics about example things
 
 ### Measurements & Fields:
 
-<optional description>
+Here you should add an optional description and links to where the user can
+get more information about the measurements.
 
 - measurement1
     - field1 (type, unit)
@@ -27,11 +30,14 @@ The example plugin gathers metrics about example things
     - tag2
 - measurement2 has the following tags:
     - tag3
-    
+
 ### Sample Queries:
 
-These are some useful queries (to generate dashboards or other) to run against data from this plugin:
+This section should contain some useful InfluxDB queries that can be used to
+get started with the plugin or to generate dashboards.  For each query listed,
+describe at a high level what data is returned.
 
+Get the max, mean, and min for the measurement in the last hour:
 ```
 SELECT max(field1), mean(field1), min(field1) FROM measurement1 WHERE tag1=bar AND time > now() - 1h GROUP BY tag
 ```
@@ -39,7 +45,7 @@ SELECT max(field1), mean(field1), min(field1) FROM measurement1 WHERE tag1=bar A
 ### Example Output:
 
 ```
-$ ./telegraf -config telegraf.conf -input-filter example -test
+$ telegraf -input-filter example -test
 measurement1,tag1=foo,tag2=bar field1=1i,field2=2.1 1453831884664956455
 measurement2,tag1=foo,tag2=bar,tag3=baz field3=1i 1453831884664956455
 ```

plugins/inputs/aerospike/aerospike.go

@@ -10,7 +10,6 @@ import (
 	"time"
 
 	"github.com/influxdata/telegraf"
-	"github.com/influxdata/telegraf/internal/errchan"
 	"github.com/influxdata/telegraf/plugins/inputs"
 
 	as "github.com/aerospike/aerospike-client-go"
@@ -41,17 +40,16 @@ func (a *Aerospike) Gather(acc telegraf.Accumulator) error {
 	}
 
 	var wg sync.WaitGroup
-	errChan := errchan.New(len(a.Servers))
 	wg.Add(len(a.Servers))
 	for _, server := range a.Servers {
 		go func(serv string) {
 			defer wg.Done()
-			errChan.C <- a.gatherServer(serv, acc)
+			acc.AddError(a.gatherServer(serv, acc))
 		}(server)
 	}
 
 	wg.Wait()
-	return errChan.Error()
+	return nil
 }
 
 func (a *Aerospike) gatherServer(hostport string, acc telegraf.Accumulator) error {

plugins/inputs/aerospike/aerospike_test.go

@@ -19,7 +19,7 @@ func TestAerospikeStatistics(t *testing.T) {
 
 	var acc testutil.Accumulator
 
-	err := a.Gather(&acc)
+	err := acc.GatherError(a.Gather)
 	require.NoError(t, err)
 
 	assert.True(t, acc.HasMeasurement("aerospike_node"))
@@ -41,8 +41,7 @@ func TestAerospikeStatisticsPartialErr(t *testing.T) {
 
 	var acc testutil.Accumulator
 
-	err := a.Gather(&acc)
-	require.Error(t, err)
+	require.Error(t, acc.GatherError(a.Gather))
 
 	assert.True(t, acc.HasMeasurement("aerospike_node"))
 	assert.True(t, acc.HasMeasurement("aerospike_namespace"))

plugins/inputs/all/all.go

@@ -2,6 +2,7 @@ package all
 
 import (
 	_ "github.com/influxdata/telegraf/plugins/inputs/aerospike"
+	_ "github.com/influxdata/telegraf/plugins/inputs/amqp_consumer"
 	_ "github.com/influxdata/telegraf/plugins/inputs/apache"
 	_ "github.com/influxdata/telegraf/plugins/inputs/bcache"
 	_ "github.com/influxdata/telegraf/plugins/inputs/cassandra"
@@ -14,6 +15,7 @@ import (
 	_ "github.com/influxdata/telegraf/plugins/inputs/couchbase"
 	_ "github.com/influxdata/telegraf/plugins/inputs/couchdb"
 	_ "github.com/influxdata/telegraf/plugins/inputs/disque"
+	_ "github.com/influxdata/telegraf/plugins/inputs/dmcache"
 	_ "github.com/influxdata/telegraf/plugins/inputs/dns_query"
 	_ "github.com/influxdata/telegraf/plugins/inputs/docker"
 	_ "github.com/influxdata/telegraf/plugins/inputs/dovecot"
@@ -27,10 +29,13 @@ import (
 	_ "github.com/influxdata/telegraf/plugins/inputs/http_response"
 	_ "github.com/influxdata/telegraf/plugins/inputs/httpjson"
 	_ "github.com/influxdata/telegraf/plugins/inputs/influxdb"
+	_ "github.com/influxdata/telegraf/plugins/inputs/internal"
+	_ "github.com/influxdata/telegraf/plugins/inputs/interrupts"
 	_ "github.com/influxdata/telegraf/plugins/inputs/ipmi_sensor"
 	_ "github.com/influxdata/telegraf/plugins/inputs/iptables"
 	_ "github.com/influxdata/telegraf/plugins/inputs/jolokia"
 	_ "github.com/influxdata/telegraf/plugins/inputs/kafka_consumer"
+	_ "github.com/influxdata/telegraf/plugins/inputs/kapacitor"
 	_ "github.com/influxdata/telegraf/plugins/inputs/kubernetes"
 	_ "github.com/influxdata/telegraf/plugins/inputs/leofs"
 	_ "github.com/influxdata/telegraf/plugins/inputs/logparser"
@@ -65,6 +70,7 @@ import (
 	_ "github.com/influxdata/telegraf/plugins/inputs/sensors"
 	_ "github.com/influxdata/telegraf/plugins/inputs/snmp"
 	_ "github.com/influxdata/telegraf/plugins/inputs/snmp_legacy"
+	_ "github.com/influxdata/telegraf/plugins/inputs/socket_listener"
 	_ "github.com/influxdata/telegraf/plugins/inputs/sqlserver"
 	_ "github.com/influxdata/telegraf/plugins/inputs/statsd"
 	_ "github.com/influxdata/telegraf/plugins/inputs/sysstat"

plugins/inputs/amqp_consumer/README.md

@@ -0,0 +1,47 @@
+# AMQP Consumer Input Plugin
+
+This plugin provides a consumer for use with AMQP 0-9-1, a promenent implementation of this protocol being [RabbitMQ](https://www.rabbitmq.com/).
+
+Metrics are read from a topic exchange using the configured queue and binding_key.
+
+Message payload should be formatted in one of the [Telegraf Data Formats](https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md).
+
+For an introduction to AMQP see:
+- https://www.rabbitmq.com/tutorials/amqp-concepts.html
+- https://www.rabbitmq.com/getstarted.html
+
+The following defaults are known to work with RabbitMQ:
+
+```toml
+# AMQP consumer plugin
+[[inputs.amqp_consumer]]
+  ## AMQP url
+  url = "amqp://localhost:5672/influxdb"
+  ## AMQP exchange
+  exchange = "telegraf"
+  ## AMQP queue name
+  queue = "telegraf"
+  ## Binding Key
+  binding_key = "#"
+
+  ## Controls how many messages the server will try to keep on the network
+  ## for consumers before receiving delivery acks.
+  #prefetch_count = 50
+
+  ## Auth method. PLAIN and EXTERNAL are supported.
+  ## Using EXTERNAL requires enabling the rabbitmq_auth_mechanism_ssl plugin as
+  ## described here: https://www.rabbitmq.com/plugins.html
+  # auth_method = "PLAIN"
+  ## Optional SSL Config
+  # ssl_ca = "/etc/telegraf/ca.pem"
+  # ssl_cert = "/etc/telegraf/cert.pem"
+  # ssl_key = "/etc/telegraf/key.pem"
+  ## Use SSL but skip chain & host verification
+  # insecure_skip_verify = false
+
+  ## Data format to output.
+  ## Each data format has its own unique set of configuration options, read
+  ## more about them here:
+  ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_OUTPUT.md
+  data_format = "influx"
+```

plugins/inputs/amqp_consumer/amqp_consumer.go

@@ -0,0 +1,280 @@
+package amqp_consumer
+
+import (
+	"fmt"
+	"log"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/streadway/amqp"
+
+	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/internal"
+	"github.com/influxdata/telegraf/plugins/inputs"
+	"github.com/influxdata/telegraf/plugins/parsers"
+)
+
+// AMQPConsumer is the top level struct for this plugin
+type AMQPConsumer struct {
+	URL string
+	// AMQP exchange
+	Exchange string
+	// Queue Name
+	Queue string
+	// Binding Key
+	BindingKey string `toml:"binding_key"`
+
+	// Controls how many messages the server will try to keep on the network
+	// for consumers before receiving delivery acks.
+	PrefetchCount int
+
+	// AMQP Auth method
+	AuthMethod string
+	// Path to CA file
+	SSLCA string `toml:"ssl_ca"`
+	// Path to host cert file
+	SSLCert string `toml:"ssl_cert"`
+	// Path to cert key file
+	SSLKey string `toml:"ssl_key"`
+	// Use SSL but skip chain & host verification
+	InsecureSkipVerify bool
+
+	parser parsers.Parser
+	conn   *amqp.Connection
+	wg     *sync.WaitGroup
+}
+
+type externalAuth struct{}
+
+func (a *externalAuth) Mechanism() string {
+	return "EXTERNAL"
+}
+func (a *externalAuth) Response() string {
+	return fmt.Sprintf("\000")
+}
+
+const (
+	DefaultAuthMethod    = "PLAIN"
+	DefaultPrefetchCount = 50
+)
+
+func (a *AMQPConsumer) SampleConfig() string {
+	return `
+  ## AMQP url
+  url = "amqp://localhost:5672/influxdb"
+  ## AMQP exchange
+  exchange = "telegraf"
+  ## AMQP queue name
+  queue = "telegraf"
+  ## Binding Key
+  binding_key = "#"
+
+  ## Maximum number of messages server should give to the worker.
+  prefetch_count = 50
+
+  ## Auth method. PLAIN and EXTERNAL are supported
+  ## Using EXTERNAL requires enabling the rabbitmq_auth_mechanism_ssl plugin as
+  ## described here: https://www.rabbitmq.com/plugins.html
+  # auth_method = "PLAIN"
+
+  ## Optional SSL Config
+  # ssl_ca = "/etc/telegraf/ca.pem"
+  # ssl_cert = "/etc/telegraf/cert.pem"
+  # ssl_key = "/etc/telegraf/key.pem"
+  ## Use SSL but skip chain & host verification
+  # insecure_skip_verify = false
+
+  ## Data format to output.
+  ## Each data format has its own unique set of configuration options, read
+  ## more about them here:
+  ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_OUTPUT.md
+  data_format = "influx"
+`
+}
+
+func (a *AMQPConsumer) Description() string {
+	return "AMQP consumer plugin"
+}
+
+func (a *AMQPConsumer) SetParser(parser parsers.Parser) {
+	a.parser = parser
+}
+
+// All gathering is done in the Start function
+func (a *AMQPConsumer) Gather(_ telegraf.Accumulator) error {
+	return nil
+}
+
+func (a *AMQPConsumer) createConfig() (*amqp.Config, error) {
+	// make new tls config
+	tls, err := internal.GetTLSConfig(
+		a.SSLCert, a.SSLKey, a.SSLCA, a.InsecureSkipVerify)
+	if err != nil {
+		return nil, err
+	}
+
+	// parse auth method
+	var sasl []amqp.Authentication // nil by default
+
+	if strings.ToUpper(a.AuthMethod) == "EXTERNAL" {
+		sasl = []amqp.Authentication{&externalAuth{}}
+	}
+
+	config := amqp.Config{
+		TLSClientConfig: tls,
+		SASL:            sasl, // if nil, it will be PLAIN
+	}
+	return &config, nil
+}
+
+// Start satisfies the telegraf.ServiceInput interface
+func (a *AMQPConsumer) Start(acc telegraf.Accumulator) error {
+	amqpConf, err := a.createConfig()
+	if err != nil {
+		return err
+	}
+
+	msgs, err := a.connect(amqpConf)
+	if err != nil {
+		return err
+	}
+
+	a.wg = &sync.WaitGroup{}
+	a.wg.Add(1)
+	go a.process(msgs, acc)
+
+	go func() {
+		err := <-a.conn.NotifyClose(make(chan *amqp.Error))
+		if err == nil {
+			return
+		}
+
+		log.Printf("I! AMQP consumer connection closed: %s; trying to reconnect", err)
+		for {
+			msgs, err := a.connect(amqpConf)
+			if err != nil {
+				log.Printf("E! AMQP connection failed: %s", err)
+				time.Sleep(10 * time.Second)
+				continue
+			}
+
+			a.wg.Add(1)
+			go a.process(msgs, acc)
+			break
+		}
+	}()
+
+	return nil
+}
+
+func (a *AMQPConsumer) connect(amqpConf *amqp.Config) (<-chan amqp.Delivery, error) {
+	conn, err := amqp.DialConfig(a.URL, *amqpConf)
+	if err != nil {
+		return nil, err
+	}
+	a.conn = conn
+
+	ch, err := conn.Channel()
+	if err != nil {
+		return nil, fmt.Errorf("Failed to open a channel: %s", err)
+	}
+
+	err = ch.ExchangeDeclare(
+		a.Exchange, // name
+		"topic",    // type
+		true,       // durable
+		false,      // auto-deleted
+		false,      // internal
+		false,      // no-wait
+		nil,        // arguments
+	)
+	if err != nil {
+		return nil, fmt.Errorf("Failed to declare an exchange: %s", err)
+	}
+
+	q, err := ch.QueueDeclare(
+		a.Queue, // queue
+		true,    // durable
+		false,   // delete when unused
+		false,   // exclusive
+		false,   // no-wait
+		nil,     // arguments
+	)
+	if err != nil {
+		return nil, fmt.Errorf("Failed to declare a queue: %s", err)
+	}
+
+	err = ch.QueueBind(
+		q.Name,       // queue
+		a.BindingKey, // binding-key
+		a.Exchange,   // exchange
+		false,
+		nil,
+	)
+	if err != nil {
+		return nil, fmt.Errorf("Failed to bind a queue: %s", err)
+	}
+
+	err = ch.Qos(
+		a.PrefetchCount,
+		0,     // prefetch-size
+		false, // global
+	)
+	if err != nil {
+		return nil, fmt.Errorf("Failed to set QoS: %s", err)
+	}
+
+	msgs, err := ch.Consume(
+		q.Name, // queue
+		"",     // consumer
+		false,  // auto-ack
+		false,  // exclusive
+		false,  // no-local
+		false,  // no-wait
+		nil,    // arguments
+	)
+	if err != nil {
+		return nil, fmt.Errorf("Failed establishing connection to queue: %s", err)
+	}
+
+	log.Println("I! Started AMQP consumer")
+	return msgs, err
+}
+
+// Read messages from queue and add them to the Accumulator
+func (a *AMQPConsumer) process(msgs <-chan amqp.Delivery, acc telegraf.Accumulator) {
+	defer a.wg.Done()
+	for d := range msgs {
+		metrics, err := a.parser.Parse(d.Body)
+		if err != nil {
+			log.Printf("E! %v: error parsing metric - %v", err, string(d.Body))
+		} else {
+			for _, m := range metrics {
+				acc.AddFields(m.Name(), m.Fields(), m.Tags(), m.Time())
+			}
+		}
+
+		d.Ack(false)
+	}
+	log.Printf("I! AMQP consumer queue closed")
+}
+
+func (a *AMQPConsumer) Stop() {
+	err := a.conn.Close()
+	if err != nil && err != amqp.ErrClosed {
+		log.Printf("E! Error closing AMQP connection: %s", err)
+		return
+	}
+	a.wg.Wait()
+	log.Println("I! Stopped AMQP service")
+}
+
+func init() {
+	inputs.Add("amqp_consumer", func() telegraf.Input {
+		return &AMQPConsumer{
+			AuthMethod:    DefaultAuthMethod,
+			PrefetchCount: DefaultPrefetchCount,
+		}
+	})
+}

plugins/inputs/apache/README.md

@@ -2,6 +2,16 @@
 
 #### Plugin arguments:
 - **urls** []string: List of apache-status URLs to collect from. Default is "http://localhost/server-status?auto".
+- **username** string: Username for HTTP basic authentication
+- **password** string: Password for HTTP basic authentication
+- **timeout** duration: time that the HTTP connection will remain waiting for response. Default 4 seconds ("4s")
+
+##### Optional SSL Config
+
+- **ssl_ca** string: the full path for the SSL CA certicate
+- **ssl_cert** string: the full path for the SSL certificate
+- **ssl_key** string: the full path for the key file
+- **insecure_skip_verify** bool: if true HTTP client will skip all SSL verifications related to peer and host. Default to false
 
 #### Description
 

plugins/inputs/apache/apache.go

@@ -8,20 +8,48 @@ import (
 	"net/url"
 	"strconv"
 	"strings"
+	"sync"
 	"time"
 
 	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/internal"
 	"github.com/influxdata/telegraf/plugins/inputs"
 )
 
 type Apache struct {
-	Urls []string
+	Urls            []string
+	Username        string
+	Password        string
+	ResponseTimeout internal.Duration
+	// Path to CA file
+	SSLCA string `toml:"ssl_ca"`
+	// Path to host cert file
+	SSLCert string `toml:"ssl_cert"`
+	// Path to cert key file
+	SSLKey string `toml:"ssl_key"`
+	// Use SSL but skip chain & host verification
+	InsecureSkipVerify bool
+
+	client *http.Client
 }
 
 var sampleConfig = `
   ## An array of Apache status URI to gather stats.
   ## Default is "http://localhost/server-status?auto".
   urls = ["http://localhost/server-status?auto"]
+  ## user credentials for basic HTTP authentication
+  username = "myuser"
+  password = "mypassword"
+
+  ## Timeout to the complete conection and reponse time in seconds
+  response_timeout = "25s" ## default to 5 seconds
+
+  ## Optional SSL Config
+  # ssl_ca = "/etc/telegraf/ca.pem"
+  # ssl_cert = "/etc/telegraf/cert.pem"
+  # ssl_key = "/etc/telegraf/key.pem"
+  ## Use SSL but skip chain & host verification
+  # insecure_skip_verify = false
 `
 
 func (n *Apache) SampleConfig() string {
@@ -36,46 +64,70 @@ func (n *Apache) Gather(acc telegraf.Accumulator) error {
 	if len(n.Urls) == 0 {
 		n.Urls = []string{"http://localhost/server-status?auto"}
 	}
+	if n.ResponseTimeout.Duration < time.Second {
+		n.ResponseTimeout.Duration = time.Second * 5
+	}
 
-	var outerr error
-	var errch = make(chan error)
+	if n.client == nil {
+		client, err := n.createHttpClient()
+		if err != nil {
+			return err
+		}
+		n.client = client
+	}
 
+	var wg sync.WaitGroup
+	wg.Add(len(n.Urls))
 	for _, u := range n.Urls {
 		addr, err := url.Parse(u)
 		if err != nil {
-			return fmt.Errorf("Unable to parse address '%s': %s", u, err)
+			acc.AddError(fmt.Errorf("Unable to parse address '%s': %s", u, err))
+			continue
 		}
 
 		go func(addr *url.URL) {
-			errch <- n.gatherUrl(addr, acc)
+			defer wg.Done()
+			acc.AddError(n.gatherUrl(addr, acc))
 		}(addr)
 	}
 
-	// Drain channel, waiting for all requests to finish and save last error.
-	for range n.Urls {
-		if err := <-errch; err != nil {
-			outerr = err
-		}
+	wg.Wait()
+	return nil
+}
+
+func (n *Apache) createHttpClient() (*http.Client, error) {
+	tlsCfg, err := internal.GetTLSConfig(
+		n.SSLCert, n.SSLKey, n.SSLCA, n.InsecureSkipVerify)
+	if err != nil {
+		return nil, err
 	}
 
-	return outerr
-}
+	client := &http.Client{
+		Transport: &http.Transport{
+			TLSClientConfig: tlsCfg,
+		},
+		Timeout: n.ResponseTimeout.Duration,
+	}
 
-var tr = &http.Transport{
-	ResponseHeaderTimeout: time.Duration(3 * time.Second),
-}
-
-var client = &http.Client{
-	Transport: tr,
-	Timeout:   time.Duration(4 * time.Second),
+	return client, nil
 }
 
 func (n *Apache) gatherUrl(addr *url.URL, acc telegraf.Accumulator) error {
-	resp, err := client.Get(addr.String())
+	req, err := http.NewRequest("GET", addr.String(), nil)
 	if err != nil {
-		return fmt.Errorf("error making HTTP request to %s: %s", addr.String(), err)
+		return fmt.Errorf("error on new request to %s : %s\n", addr.String(), err)
+	}
+
+	if len(n.Username) != 0 && len(n.Password) != 0 {
+		req.SetBasicAuth(n.Username, n.Password)
+	}
+
+	resp, err := n.client.Do(req)
+	if err != nil {
+		return fmt.Errorf("error on request to %s : %s\n", addr.String(), err)
 	}
 	defer resp.Body.Close()
+
 	if resp.StatusCode != http.StatusOK {
 		return fmt.Errorf("%s returned HTTP status %s", addr.String(), resp.Status)
 	}

plugins/inputs/apache/apache_test.go

@@ -41,7 +41,7 @@ func TestHTTPApache(t *testing.T) {
 	}
 
 	var acc testutil.Accumulator
-	err := a.Gather(&acc)
+	err := acc.GatherError(a.Gather)
 	require.NoError(t, err)
 
 	fields := map[string]interface{}{

plugins/inputs/cassandra/cassandra.go

@@ -7,7 +7,6 @@ import (
 	"github.com/influxdata/telegraf"
 	"github.com/influxdata/telegraf/plugins/inputs"
 	"io/ioutil"
-	"log"
 	"net/http"
 	"net/url"
 	"strings"
@@ -123,8 +122,8 @@ func (j javaMetric) addTagsFields(out map[string]interface{}) {
 		}
 		j.acc.AddFields(tokens["class"]+tokens["type"], fields, tags)
 	} else {
-		fmt.Printf("Missing key 'value' in '%s' output response\n%v\n",
-			j.metric, out)
+		j.acc.AddError(fmt.Errorf("Missing key 'value' in '%s' output response\n%v\n",
+			j.metric, out))
 	}
 }
 
@@ -155,8 +154,8 @@ func (c cassandraMetric) addTagsFields(out map[string]interface{}) {
 				addCassandraMetric(k, c, v.(map[string]interface{}))
 			}
 		} else {
-			fmt.Printf("Missing key 'value' in '%s' output response\n%v\n",
-				c.metric, out)
+			c.acc.AddError(fmt.Errorf("Missing key 'value' in '%s' output response\n%v\n",
+				c.metric, out))
 			return
 		}
 	} else {
@@ -164,8 +163,8 @@ func (c cassandraMetric) addTagsFields(out map[string]interface{}) {
 			addCassandraMetric(r.(map[string]interface{})["mbean"].(string),
 				c, values.(map[string]interface{}))
 		} else {
-			fmt.Printf("Missing key 'value' in '%s' output response\n%v\n",
-				c.metric, out)
+			c.acc.AddError(fmt.Errorf("Missing key 'value' in '%s' output response\n%v\n",
+				c.metric, out))
 			return
 		}
 	}
@@ -274,8 +273,8 @@ func (c *Cassandra) Gather(acc telegraf.Accumulator) error {
 				m = newCassandraMetric(serverTokens["host"], metric, acc)
 			} else {
 				// unsupported metric type
-				log.Printf("I! Unsupported Cassandra metric [%s], skipping",
-					metric)
+				acc.AddError(fmt.Errorf("E! Unsupported Cassandra metric [%s], skipping",
+					metric))
 				continue
 			}
 
@@ -283,17 +282,21 @@ func (c *Cassandra) Gather(acc telegraf.Accumulator) error {
 			requestUrl, err := url.Parse("http://" + serverTokens["host"] + ":" +
 				serverTokens["port"] + context + metric)
 			if err != nil {
-				return err
+				acc.AddError(err)
+				continue
 			}
 			if serverTokens["user"] != "" && serverTokens["passwd"] != "" {
 				requestUrl.User = url.UserPassword(serverTokens["user"],
 					serverTokens["passwd"])
 			}
-			fmt.Printf("host %s url %s\n", serverTokens["host"], requestUrl)
 
 			out, err := c.getAttr(requestUrl)
+			if err != nil {
+				acc.AddError(err)
+				continue
+			}
 			if out["status"] != 200.0 {
-				fmt.Printf("URL returned with status %v\n", out["status"])
+				acc.AddError(fmt.Errorf("URL returned with status %v\n", out["status"]))
 				continue
 			}
 			m.addTagsFields(out)

plugins/inputs/cassandra/cassandra_test.go

@@ -151,7 +151,7 @@ func TestHttpJsonJavaMultiValue(t *testing.T) {
 
 	var acc testutil.Accumulator
 	acc.SetDebug(true)
-	err := cassandra.Gather(&acc)
+	err := acc.GatherError(cassandra.Gather)
 
 	assert.Nil(t, err)
 	assert.Equal(t, 2, len(acc.Metrics))
@@ -180,7 +180,7 @@ func TestHttpJsonJavaMultiType(t *testing.T) {
 
 	var acc testutil.Accumulator
 	acc.SetDebug(true)
-	err := cassandra.Gather(&acc)
+	err := acc.GatherError(cassandra.Gather)
 
 	assert.Nil(t, err)
 	assert.Equal(t, 2, len(acc.Metrics))
@@ -197,16 +197,17 @@ func TestHttpJsonJavaMultiType(t *testing.T) {
 }
 
 // Test that the proper values are ignored or collected
-func TestHttpJsonOn404(t *testing.T) {
+func TestHttp404(t *testing.T) {
 
-	jolokia := genJolokiaClientStub(validJavaMultiValueJSON, 404, Servers,
+	jolokia := genJolokiaClientStub(invalidJSON, 404, Servers,
 		[]string{HeapMetric})
 
 	var acc testutil.Accumulator
-	err := jolokia.Gather(&acc)
+	err := acc.GatherError(jolokia.Gather)
 
-	assert.Nil(t, err)
+	assert.Error(t, err)
 	assert.Equal(t, 0, len(acc.Metrics))
+	assert.Contains(t, err.Error(), "has status code 404")
 }
 
 // Test that the proper values are ignored or collected for class=Cassandra
@@ -214,7 +215,7 @@ func TestHttpJsonCassandraMultiValue(t *testing.T) {
 	cassandra := genJolokiaClientStub(validCassandraMultiValueJSON, 200, Servers, []string{ReadLatencyMetric})
 
 	var acc testutil.Accumulator
-	err := cassandra.Gather(&acc)
+	err := acc.GatherError(cassandra.Gather)
 
 	assert.Nil(t, err)
 	assert.Equal(t, 1, len(acc.Metrics))
@@ -246,7 +247,7 @@ func TestHttpJsonCassandraNestedMultiValue(t *testing.T) {
 
 	var acc testutil.Accumulator
 	acc.SetDebug(true)
-	err := cassandra.Gather(&acc)
+	err := acc.GatherError(cassandra.Gather)
 
 	assert.Nil(t, err)
 	assert.Equal(t, 2, len(acc.Metrics))

plugins/inputs/ceph/README.md

@@ -82,7 +82,7 @@ the cluster.  The currently supported commands are:
 
   ## Whether to gather statistics via ceph commands, requires ceph_user and ceph_config
   ## to be specified
-  gather_cluster_stats = true
+  gather_cluster_stats = false
 ```
 
 ### Measurements & Fields:
@@ -117,7 +117,7 @@ All fields are collected under the **ceph** measurement and stored as float64s.
   * recovering\_objects\_per\_sec (float)
 
 * ceph\_pgmap\_state
-  * state name e.g. active+clean (float)
+  * count (float)
 
 * ceph\_usage
   * bytes\_used (float)
@@ -186,7 +186,7 @@ All measurements will have the following tags:
 
 *Cluster Stats*
 
-* ceph\_pg\_state has the following tags:
+* ceph\_pgmap\_state has the following tags:
   * state (state for which the value applies e.g. active+clean, active+remapped+backfill)
 * ceph\_pool\_usage has the following tags:
   * id
@@ -213,7 +213,8 @@ telegraf -test -config /etc/telegraf/telegraf.conf -config-directory /etc/telegr
 <pre>
 > ceph_osdmap,host=ceph-mon-0 epoch=170772,full=false,nearfull=false,num_in_osds=340,num_osds=340,num_remapped_pgs=0,num_up_osds=340 1468841037000000000
 > ceph_pgmap,host=ceph-mon-0 bytes_avail=634895531270144,bytes_total=812117151809536,bytes_used=177221620539392,data_bytes=56979991615058,num_pgs=22952,op_per_sec=15869,read_bytes_sec=43956026,version=39387592,write_bytes_sec=165344818 1468841037000000000
-> ceph_pgmap_state,host=ceph-mon-0 active+clean=22952 1468928660000000000
+> ceph_pgmap_state,host=ceph-mon-0,state=active+clean count=22952 1468928660000000000
+> ceph_pgmap_state,host=ceph-mon-0,state=active+degraded count=16 1468928660000000000
 > ceph_usage,host=ceph-mon-0 total_avail_bytes=634895514791936,total_bytes=812117151809536,total_used_bytes=177221637017600 1468841037000000000
 > ceph_pool_usage,host=ceph-mon-0,id=150,name=cinder.volumes bytes_used=12648553794802,kb_used=12352103316,max_avail=154342562489244,objects=3026295 1468841037000000000
 > ceph_pool_usage,host=ceph-mon-0,id=182,name=cinder.volumes.flash bytes_used=8541308223964,kb_used=8341121313,max_avail=39388593563936,objects=2075066 1468841037000000000

plugins/inputs/ceph/ceph.go

@@ -4,13 +4,14 @@ import (
 	"bytes"
 	"encoding/json"
 	"fmt"
-	"github.com/influxdata/telegraf"
-	"github.com/influxdata/telegraf/plugins/inputs"
 	"io/ioutil"
 	"log"
 	"os/exec"
 	"path/filepath"
 	"strings"
+
+	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/plugins/inputs"
 )
 
 const (
@@ -68,7 +69,7 @@ var sampleConfig = `
   gather_admin_socket_stats = true
 
   ## Whether to gather statistics via ceph commands
-  gather_cluster_stats = true
+  gather_cluster_stats = false
 `
 
 func (c *Ceph) SampleConfig() string {
@@ -100,15 +101,15 @@ func (c *Ceph) gatherAdminSocketStats(acc telegraf.Accumulator) error {
 	for _, s := range sockets {
 		dump, err := perfDump(c.CephBinary, s)
 		if err != nil {
-			log.Printf("E! error reading from socket '%s': %v", s.socket, err)
+			acc.AddError(fmt.Errorf("E! error reading from socket '%s': %v", s.socket, err))
 			continue
 		}
 		data, err := parseDump(dump)
 		if err != nil {
-			log.Printf("E! error parsing dump from socket '%s': %v", s.socket, err)
+			acc.AddError(fmt.Errorf("E! error parsing dump from socket '%s': %v", s.socket, err))
 			continue
 		}
-		for tag, metrics := range *data {
+		for tag, metrics := range data {
 			acc.AddFields(measurement,
 				map[string]interface{}(metrics),
 				map[string]string{"type": s.sockType, "id": s.sockId, "collection": tag})
@@ -244,25 +245,19 @@ type taggedMetricMap map[string]metricMap
 
 // Parses a raw JSON string into a taggedMetricMap
 // Delegates the actual parsing to newTaggedMetricMap(..)
-func parseDump(dump string) (*taggedMetricMap, error) {
+func parseDump(dump string) (taggedMetricMap, error) {
 	data := make(map[string]interface{})
 	err := json.Unmarshal([]byte(dump), &data)
 	if err != nil {
 		return nil, fmt.Errorf("failed to parse json: '%s': %v", dump, err)
 	}
 
-	tmm := newTaggedMetricMap(data)
-
-	if err != nil {
-		return nil, fmt.Errorf("failed to tag dataset: '%v': %v", tmm, err)
-	}
-
-	return tmm, nil
+	return newTaggedMetricMap(data), nil
 }
 
 // Builds a TaggedMetricMap out of a generic string map.
 // The top-level key is used as a tag and all sub-keys are flattened into metrics
-func newTaggedMetricMap(data map[string]interface{}) *taggedMetricMap {
+func newTaggedMetricMap(data map[string]interface{}) taggedMetricMap {
 	tmm := make(taggedMetricMap)
 	for tag, datapoints := range data {
 		mm := make(metricMap)
@@ -271,7 +266,7 @@ func newTaggedMetricMap(data map[string]interface{}) *taggedMetricMap {
 		}
 		tmm[tag] = mm
 	}
-	return &tmm
+	return tmm
 }
 
 // Recursively flattens any k-v hierarchy present in data.
@@ -376,36 +371,53 @@ func decodeStatusPgmap(acc telegraf.Accumulator, data map[string]interface{}) er
 	return nil
 }
 
-func decodeStatusPgmapState(acc telegraf.Accumulator, data map[string]interface{}) error {
+func extractPgmapStates(data map[string]interface{}) ([]interface{}, error) {
+	const key = "pgs_by_state"
+
 	pgmap, ok := data["pgmap"].(map[string]interface{})
 	if !ok {
-		return fmt.Errorf("WARNING %s - unable to decode pgmap", measurement)
+		return nil, fmt.Errorf("WARNING %s - unable to decode pgmap", measurement)
 	}
-	fields := make(map[string]interface{})
-	for key, value := range pgmap {
-		switch value.(type) {
-		case []interface{}:
-			if key != "pgs_by_state" {
-				continue
-			}
-			for _, state := range value.([]interface{}) {
-				state_map, ok := state.(map[string]interface{})
-				if !ok {
-					return fmt.Errorf("WARNING %s - unable to decode pg state", measurement)
-				}
-				state_name, ok := state_map["state_name"].(string)
-				if !ok {
-					return fmt.Errorf("WARNING %s - unable to decode pg state name", measurement)
-				}
-				state_count, ok := state_map["count"].(float64)
-				if !ok {
-					return fmt.Errorf("WARNING %s - unable to decode pg state count", measurement)
-				}
-				fields[state_name] = state_count
-			}
+
+	s, ok := pgmap[key]
+	if !ok {
+		return nil, fmt.Errorf("WARNING %s - pgmap is missing the %s field", measurement, key)
+	}
+
+	states, ok := s.([]interface{})
+	if !ok {
+		return nil, fmt.Errorf("WARNING %s - pgmap[%s] is not a list", measurement, key)
+	}
+	return states, nil
+}
+
+func decodeStatusPgmapState(acc telegraf.Accumulator, data map[string]interface{}) error {
+	states, err := extractPgmapStates(data)
+	if err != nil {
+		return err
+	}
+	for _, state := range states {
+		stateMap, ok := state.(map[string]interface{})
+		if !ok {
+			return fmt.Errorf("WARNING %s - unable to decode pg state", measurement)
 		}
+		stateName, ok := stateMap["state_name"].(string)
+		if !ok {
+			return fmt.Errorf("WARNING %s - unable to decode pg state name", measurement)
+		}
+		stateCount, ok := stateMap["count"].(float64)
+		if !ok {
+			return fmt.Errorf("WARNING %s - unable to decode pg state count", measurement)
+		}
+
+		tags := map[string]string{
+			"state": stateName,
+		}
+		fields := map[string]interface{}{
+			"count": stateCount,
+		}
+		acc.AddFields("ceph_pgmap_state", fields, tags)
 	}
-	acc.AddFields("ceph_pgmap_state", fields, map[string]string{})
 	return nil
 }
 

plugins/inputs/ceph/ceph_test.go

@@ -1,15 +1,17 @@
 package ceph
 
 import (
+	"encoding/json"
 	"fmt"
-	"github.com/influxdata/telegraf/testutil"
-	"github.com/stretchr/testify/assert"
 	"io/ioutil"
 	"os"
 	"path"
 	"strconv"
 	"strings"
 	"testing"
+
+	"github.com/influxdata/telegraf/testutil"
+	"github.com/stretchr/testify/assert"
 )
 
 const (
@@ -24,15 +26,38 @@ func TestParseSockId(t *testing.T) {
 func TestParseMonDump(t *testing.T) {
 	dump, err := parseDump(monPerfDump)
 	assert.NoError(t, err)
-	assert.InEpsilon(t, 5678670180, (*dump)["cluster"]["osd_kb_used"], epsilon)
-	assert.InEpsilon(t, 6866.540527000, (*dump)["paxos"]["store_state_latency.sum"], epsilon)
+	assert.InEpsilon(t, 5678670180, dump["cluster"]["osd_kb_used"], epsilon)
+	assert.InEpsilon(t, 6866.540527000, dump["paxos"]["store_state_latency.sum"], epsilon)
 }
 
 func TestParseOsdDump(t *testing.T) {
 	dump, err := parseDump(osdPerfDump)
 	assert.NoError(t, err)
-	assert.InEpsilon(t, 552132.109360000, (*dump)["filestore"]["commitcycle_interval.sum"], epsilon)
-	assert.Equal(t, float64(0), (*dump)["mutex-FileJournal::finisher_lock"]["wait.avgcount"])
+	assert.InEpsilon(t, 552132.109360000, dump["filestore"]["commitcycle_interval.sum"], epsilon)
+	assert.Equal(t, float64(0), dump["mutex-FileJournal::finisher_lock"]["wait.avgcount"])
+}
+
+func TestDecodeStatusPgmapState(t *testing.T) {
+	data := make(map[string]interface{})
+	err := json.Unmarshal([]byte(clusterStatusDump), &data)
+	assert.NoError(t, err)
+
+	acc := &testutil.Accumulator{}
+	err = decodeStatusPgmapState(acc, data)
+	assert.NoError(t, err)
+
+	var results = []struct {
+		fields map[string]interface{}
+		tags   map[string]string
+	}{
+		{map[string]interface{}{"count": float64(2560)}, map[string]string{"state": "active+clean"}},
+		{map[string]interface{}{"count": float64(10)}, map[string]string{"state": "active+scrubbing"}},
+		{map[string]interface{}{"count": float64(5)}, map[string]string{"state": "active+backfilling"}},
+	}
+
+	for _, r := range results {
+		acc.AssertContainsTaggedFields(t, "ceph_pgmap_state", r.fields, r.tags)
+	}
 }
 
 func TestGather(t *testing.T) {
@@ -685,3 +710,127 @@ var osdPerfDump = `
       "wait": { "avgcount": 0,
           "sum": 0.000000000}}}
 `
+var clusterStatusDump = `
+{
+  "health": {
+    "health": {
+      "health_services": [
+        {
+          "mons": [
+            {
+              "name": "a",
+              "kb_total": 114289256,
+              "kb_used": 26995516,
+              "kb_avail": 81465132,
+              "avail_percent": 71,
+              "last_updated": "2017-01-03 17:20:57.595004",
+              "store_stats": {
+                "bytes_total": 942117141,
+                "bytes_sst": 0,
+                "bytes_log": 4345406,
+                "bytes_misc": 937771735,
+                "last_updated": "0.000000"
+              },
+              "health": "HEALTH_OK"
+            },
+            {
+              "name": "b",
+              "kb_total": 114289256,
+              "kb_used": 27871624,
+              "kb_avail": 80589024,
+              "avail_percent": 70,
+              "last_updated": "2017-01-03 17:20:47.784331",
+              "store_stats": {
+                "bytes_total": 454853104,
+                "bytes_sst": 0,
+                "bytes_log": 5788320,
+                "bytes_misc": 449064784,
+                "last_updated": "0.000000"
+              },
+              "health": "HEALTH_OK"
+            },
+            {
+              "name": "c",
+              "kb_total": 130258508,
+              "kb_used": 38076996,
+              "kb_avail": 85541692,
+              "avail_percent": 65,
+              "last_updated": "2017-01-03 17:21:03.311123",
+              "store_stats": {
+                "bytes_total": 455555199,
+                "bytes_sst": 0,
+                "bytes_log": 6950876,
+                "bytes_misc": 448604323,
+                "last_updated": "0.000000"
+              },
+              "health": "HEALTH_OK"
+            }
+          ]
+        }
+      ]
+    },
+    "timechecks": {
+      "epoch": 504,
+      "round": 34642,
+      "round_status": "finished",
+      "mons": [
+        { "name": "a", "skew": 0, "latency": 0, "health": "HEALTH_OK" },
+        { "name": "b", "skew": -0, "latency": 0.000951, "health": "HEALTH_OK" },
+        { "name": "c", "skew": -0, "latency": 0.000946, "health": "HEALTH_OK" }
+      ]
+    },
+    "summary": [],
+    "overall_status": "HEALTH_OK",
+    "detail": []
+  },
+  "fsid": "01234567-abcd-9876-0123-ffeeddccbbaa",
+  "election_epoch": 504,
+  "quorum": [ 0, 1, 2 ],
+  "quorum_names": [ "a", "b", "c" ],
+  "monmap": {
+    "epoch": 17,
+    "fsid": "01234567-abcd-9876-0123-ffeeddccbbaa",
+    "modified": "2016-04-11 14:01:52.600198",
+    "created": "0.000000",
+    "mons": [
+      { "rank": 0, "name": "a", "addr": "192.168.0.1:6789/0" },
+      { "rank": 1, "name": "b", "addr": "192.168.0.2:6789/0" },
+      { "rank": 2, "name": "c", "addr": "192.168.0.3:6789/0" }
+    ]
+  },
+  "osdmap": {
+    "osdmap": {
+      "epoch": 21734,
+      "num_osds": 24,
+      "num_up_osds": 24,
+      "num_in_osds": 24,
+      "full": false,
+      "nearfull": false,
+      "num_remapped_pgs": 0
+    }
+  },
+  "pgmap": {
+    "pgs_by_state": [
+      { "state_name": "active+clean", "count": 2560 },
+      { "state_name": "active+scrubbing", "count": 10 },
+      { "state_name": "active+backfilling", "count": 5 }
+    ],
+    "version": 52314277,
+    "num_pgs": 2560,
+    "data_bytes": 2700031960713,
+    "bytes_used": 7478347665408,
+    "bytes_avail": 9857462382592,
+    "bytes_total": 17335810048000,
+    "read_bytes_sec": 0,
+    "write_bytes_sec": 367217,
+    "op_per_sec": 98
+  },
+  "mdsmap": {
+    "epoch": 1,
+    "up": 0,
+    "in": 0,
+    "max": 0,
+    "by_rank": []
+  }
+}
+`

plugins/inputs/cgroup/cgroup_linux.go

@@ -22,10 +22,11 @@ func (g *CGroup) Gather(acc telegraf.Accumulator) error {
 
 	for dir := range list {
 		if dir.err != nil {
-			return dir.err
+			acc.AddError(dir.err)
+			continue
 		}
 		if err := g.gatherDir(dir.path, acc); err != nil {
-			return err
+			acc.AddError(err)
 		}
 	}
 

plugins/inputs/cgroup/cgroup_test.go

@@ -24,7 +24,7 @@ var cg1 = &CGroup{
 func TestCgroupStatistics_1(t *testing.T) {
 	var acc testutil.Accumulator
 
-	err := cg1.Gather(&acc)
+	err := acc.GatherError(cg1.Gather)
 	require.NoError(t, err)
 
 	tags := map[string]string{
@@ -56,7 +56,7 @@ var cg2 = &CGroup{
 func TestCgroupStatistics_2(t *testing.T) {
 	var acc testutil.Accumulator
 
-	err := cg2.Gather(&acc)
+	err := acc.GatherError(cg2.Gather)
 	require.NoError(t, err)
 
 	tags := map[string]string{
@@ -81,7 +81,7 @@ var cg3 = &CGroup{
 func TestCgroupStatistics_3(t *testing.T) {
 	var acc testutil.Accumulator
 
-	err := cg3.Gather(&acc)
+	err := acc.GatherError(cg3.Gather)
 	require.NoError(t, err)
 
 	tags := map[string]string{
@@ -108,7 +108,7 @@ var cg4 = &CGroup{
 func TestCgroupStatistics_4(t *testing.T) {
 	var acc testutil.Accumulator
 
-	err := cg4.Gather(&acc)
+	err := acc.GatherError(cg4.Gather)
 	require.NoError(t, err)
 
 	tags := map[string]string{
@@ -140,7 +140,7 @@ var cg5 = &CGroup{
 func TestCgroupStatistics_5(t *testing.T) {
 	var acc testutil.Accumulator
 
-	err := cg5.Gather(&acc)
+	err := acc.GatherError(cg5.Gather)
 	require.NoError(t, err)
 
 	tags := map[string]string{
@@ -167,7 +167,7 @@ var cg6 = &CGroup{
 func TestCgroupStatistics_6(t *testing.T) {
 	var acc testutil.Accumulator
 
-	err := cg6.Gather(&acc)
+	err := acc.GatherError(cg6.Gather)
 	require.NoError(t, err)
 
 	tags := map[string]string{

plugins/inputs/cloudwatch/README.md

@@ -42,9 +42,10 @@ API endpoint. In the following order the plugin will attempt to authenticate.
   namespace = "AWS/ELB"
 
   ## Maximum requests per second. Note that the global default AWS rate limit is
-  ## 10 reqs/sec, so if you define multiple namespaces, these should add up to a
-  ## maximum of 10. Optional - default value is 10.
-  ratelimit = 10
+  ## 400 reqs/sec, so if you define multiple namespaces, these should add up to a
+  ## maximum of 400. Optional - default value is 200.
+  ## See http://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/cloudwatch_limits.html
+  ratelimit = 200
 
   ## Metrics to Pull (optional)
   ## Defaults to all Metrics in Namespace if nothing is provided

plugins/inputs/cloudwatch/cloudwatch.go

@@ -13,7 +13,6 @@ import (
 	"github.com/influxdata/telegraf"
 	"github.com/influxdata/telegraf/internal"
 	internalaws "github.com/influxdata/telegraf/internal/config/aws"
-	"github.com/influxdata/telegraf/internal/errchan"
 	"github.com/influxdata/telegraf/internal/limiter"
 	"github.com/influxdata/telegraf/plugins/inputs"
 )
@@ -105,9 +104,10 @@ func (c *CloudWatch) SampleConfig() string {
   namespace = "AWS/ELB"
 
   ## Maximum requests per second. Note that the global default AWS rate limit is
-  ## 10 reqs/sec, so if you define multiple namespaces, these should add up to a
-  ## maximum of 10. Optional - default value is 10.
-  ratelimit = 10
+  ## 400 reqs/sec, so if you define multiple namespaces, these should add up to a
+  ## maximum of 400. Optional - default value is 200.
+  ## See http://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/cloudwatch_limits.html
+  ratelimit = 200
 
   ## Metrics to Pull (optional)
   ## Defaults to all Metrics in Namespace if nothing is provided
@@ -126,11 +126,7 @@ func (c *CloudWatch) Description() string {
 	return "Pull Metric Statistics from Amazon CloudWatch"
 }
 
-func (c *CloudWatch) Gather(acc telegraf.Accumulator) error {
-	if c.client == nil {
-		c.initializeCloudWatch()
-	}
-
+func SelectMetrics(c *CloudWatch) ([]*cloudwatch.Metric, error) {
 	var metrics []*cloudwatch.Metric
 
 	// check for provided metric filter
@@ -155,11 +151,11 @@ func (c *CloudWatch) Gather(acc telegraf.Accumulator) error {
 			} else {
 				allMetrics, err := c.fetchNamespaceMetrics()
 				if err != nil {
-					return err
+					return nil, err
 				}
 				for _, name := range m.MetricNames {
 					for _, metric := range allMetrics {
-						if isSelected(metric, m.Dimensions) {
+						if isSelected(name, metric, m.Dimensions) {
 							metrics = append(metrics, &cloudwatch.Metric{
 								Namespace:  aws.String(c.Namespace),
 								MetricName: aws.String(name),
@@ -169,18 +165,26 @@ func (c *CloudWatch) Gather(acc telegraf.Accumulator) error {
 					}
 				}
 			}
-
 		}
 	} else {
 		var err error
 		metrics, err = c.fetchNamespaceMetrics()
 		if err != nil {
-			return err
+			return nil, err
 		}
 	}
+	return metrics, nil
+}
 
-	metricCount := len(metrics)
-	errChan := errchan.New(metricCount)
+func (c *CloudWatch) Gather(acc telegraf.Accumulator) error {
+	if c.client == nil {
+		c.initializeCloudWatch()
+	}
+
+	metrics, err := SelectMetrics(c)
+	if err != nil {
+		return err
+	}
 
 	now := time.Now()
 
@@ -195,12 +199,12 @@ func (c *CloudWatch) Gather(acc telegraf.Accumulator) error {
 		<-lmtr.C
 		go func(inm *cloudwatch.Metric) {
 			defer wg.Done()
-			c.gatherMetric(acc, inm, now, errChan.C)
+			acc.AddError(c.gatherMetric(acc, inm, now))
 		}(m)
 	}
 	wg.Wait()
 
-	return errChan.Error()
+	return nil
 }
 
 func init() {
@@ -208,7 +212,7 @@ func init() {
 		ttl, _ := time.ParseDuration("1hr")
 		return &CloudWatch{
 			CacheTTL:  internal.Duration{Duration: ttl},
-			RateLimit: 10,
+			RateLimit: 200,
 		}
 	})
 }
@@ -278,13 +282,11 @@ func (c *CloudWatch) gatherMetric(
 	acc telegraf.Accumulator,
 	metric *cloudwatch.Metric,
 	now time.Time,
-	errChan chan error,
-) {
+) error {
 	params := c.getStatisticsInput(metric, now)
 	resp, err := c.client.GetMetricStatistics(params)
 	if err != nil {
-		errChan <- err
-		return
+		return err
 	}
 
 	for _, point := range resp.Datapoints {
@@ -319,7 +321,7 @@ func (c *CloudWatch) gatherMetric(
 		acc.AddFields(formatMeasurement(c.Namespace), fields, tags, *point.Timestamp)
 	}
 
-	errChan <- nil
+	return nil
 }
 
 /*
@@ -380,7 +382,10 @@ func hasWilcard(dimensions []*Dimension) bool {
 	return false
 }
 
-func isSelected(metric *cloudwatch.Metric, dimensions []*Dimension) bool {
+func isSelected(name string, metric *cloudwatch.Metric, dimensions []*Dimension) bool {
+	if name != *metric.MetricName {
+		return false
+	}
 	if len(metric.Dimensions) != len(dimensions) {
 		return false
 	}

plugins/inputs/cloudwatch/cloudwatch_test.go

@@ -11,9 +11,9 @@ import (
 	"github.com/stretchr/testify/assert"
 )
 
-type mockCloudWatchClient struct{}
+type mockGatherCloudWatchClient struct{}
 
-func (m *mockCloudWatchClient) ListMetrics(params *cloudwatch.ListMetricsInput) (*cloudwatch.ListMetricsOutput, error) {
+func (m *mockGatherCloudWatchClient) ListMetrics(params *cloudwatch.ListMetricsInput) (*cloudwatch.ListMetricsOutput, error) {
 	metric := &cloudwatch.Metric{
 		Namespace:  params.Namespace,
 		MetricName: aws.String("Latency"),
@@ -31,7 +31,7 @@ func (m *mockCloudWatchClient) ListMetrics(params *cloudwatch.ListMetricsInput)
 	return result, nil
 }
 
-func (m *mockCloudWatchClient) GetMetricStatistics(params *cloudwatch.GetMetricStatisticsInput) (*cloudwatch.GetMetricStatisticsOutput, error) {
+func (m *mockGatherCloudWatchClient) GetMetricStatistics(params *cloudwatch.GetMetricStatisticsInput) (*cloudwatch.GetMetricStatisticsOutput, error) {
 	dataPoint := &cloudwatch.Datapoint{
 		Timestamp:   params.EndTime,
 		Minimum:     aws.Float64(0.1),
@@ -58,13 +58,13 @@ func TestGather(t *testing.T) {
 		Namespace: "AWS/ELB",
 		Delay:     internalDuration,
 		Period:    internalDuration,
-		RateLimit: 10,
+		RateLimit: 200,
 	}
 
 	var acc testutil.Accumulator
-	c.client = &mockCloudWatchClient{}
+	c.client = &mockGatherCloudWatchClient{}
 
-	c.Gather(&acc)
+	acc.GatherError(c.Gather)
 
 	fields := map[string]interface{}{}
 	fields["latency_minimum"] = 0.1
@@ -83,6 +83,94 @@ func TestGather(t *testing.T) {
 
 }
 
+type mockSelectMetricsCloudWatchClient struct{}
+
+func (m *mockSelectMetricsCloudWatchClient) ListMetrics(params *cloudwatch.ListMetricsInput) (*cloudwatch.ListMetricsOutput, error) {
+	metrics := []*cloudwatch.Metric{}
+	// 4 metrics are available
+	metricNames := []string{"Latency", "RequestCount", "HealthyHostCount", "UnHealthyHostCount"}
+	// for 3 ELBs
+	loadBalancers := []string{"lb-1", "lb-2", "lb-3"}
+	// in 2 AZs
+	availabilityZones := []string{"us-east-1a", "us-east-1b"}
+	for _, m := range metricNames {
+		for _, lb := range loadBalancers {
+			// For each metric/ELB pair, we get an aggregate value across all AZs.
+			metrics = append(metrics, &cloudwatch.Metric{
+				Namespace:  aws.String("AWS/ELB"),
+				MetricName: aws.String(m),
+				Dimensions: []*cloudwatch.Dimension{
+					&cloudwatch.Dimension{
+						Name:  aws.String("LoadBalancerName"),
+						Value: aws.String(lb),
+					},
+				},
+			})
+			for _, az := range availabilityZones {
+				// We get a metric for each metric/ELB/AZ triplet.
+				metrics = append(metrics, &cloudwatch.Metric{
+					Namespace:  aws.String("AWS/ELB"),
+					MetricName: aws.String(m),
+					Dimensions: []*cloudwatch.Dimension{
+						&cloudwatch.Dimension{
+							Name:  aws.String("LoadBalancerName"),
+							Value: aws.String(lb),
+						},
+						&cloudwatch.Dimension{
+							Name:  aws.String("AvailabilityZone"),
+							Value: aws.String(az),
+						},
+					},
+				})
+			}
+		}
+	}
+
+	result := &cloudwatch.ListMetricsOutput{
+		Metrics: metrics,
+	}
+	return result, nil
+}
+
+func (m *mockSelectMetricsCloudWatchClient) GetMetricStatistics(params *cloudwatch.GetMetricStatisticsInput) (*cloudwatch.GetMetricStatisticsOutput, error) {
+	return nil, nil
+}
+
+func TestSelectMetrics(t *testing.T) {
+	duration, _ := time.ParseDuration("1m")
+	internalDuration := internal.Duration{
+		Duration: duration,
+	}
+	c := &CloudWatch{
+		Region:    "us-east-1",
+		Namespace: "AWS/ELB",
+		Delay:     internalDuration,
+		Period:    internalDuration,
+		RateLimit: 200,
+		Metrics: []*Metric{
+			&Metric{
+				MetricNames: []string{"Latency", "RequestCount"},
+				Dimensions: []*Dimension{
+					&Dimension{
+						Name:  "LoadBalancerName",
+						Value: "*",
+					},
+					&Dimension{
+						Name:  "AvailabilityZone",
+						Value: "*",
+					},
+				},
+			},
+		},
+	}
+	c.client = &mockSelectMetricsCloudWatchClient{}
+	metrics, err := SelectMetrics(c)
+	// We've asked for 2 (out of 4) metrics, over all 3 load balancers in all 2
+	// AZs. We should get 12 metrics.
+	assert.Equal(t, 12, len(metrics))
+	assert.Nil(t, err)
+}
+
 func TestGenerateStatisticsInputParams(t *testing.T) {
 	d := &cloudwatch.Dimension{
 		Name:  aws.String("LoadBalancerName"),
@@ -119,14 +207,13 @@ func TestGenerateStatisticsInputParams(t *testing.T) {
 }
 
 func TestMetricsCacheTimeout(t *testing.T) {
-	ttl, _ := time.ParseDuration("5ms")
 	cache := &MetricCache{
 		Metrics: []*cloudwatch.Metric{},
 		Fetched: time.Now(),
-		TTL:     ttl,
+		TTL:     time.Minute,
 	}
 
 	assert.True(t, cache.IsValid())
-	time.Sleep(ttl)
+	cache.Fetched = time.Now().Add(-time.Minute)
 	assert.False(t, cache.IsValid())
 }

plugins/inputs/conntrack/conntrack.go

@@ -11,7 +11,6 @@ import (
 
 	"github.com/influxdata/telegraf"
 	"github.com/influxdata/telegraf/plugins/inputs"
-	"log"
 	"path/filepath"
 )
 
@@ -93,15 +92,15 @@ func (c *Conntrack) Gather(acc telegraf.Accumulator) error {
 
 			contents, err := ioutil.ReadFile(fName)
 			if err != nil {
-				log.Printf("E! failed to read file '%s': %v", fName, err)
+				acc.AddError(fmt.Errorf("E! failed to read file '%s': %v", fName, err))
 				continue
 			}
 
 			v := strings.TrimSpace(string(contents))
 			fields[metricKey], err = strconv.ParseFloat(v, 64)
 			if err != nil {
-				log.Printf("E! failed to parse metric, expected number but "+
-					" found '%s': %v", v, err)
+				acc.AddError(fmt.Errorf("E! failed to parse metric, expected number but "+
+					" found '%s': %v", v, err))
 			}
 		}
 	}

plugins/inputs/consul/README.md

@@ -29,18 +29,25 @@ to query the data. It will not report the [telemetry](https://www.consul.io/docs
 Tags:
 - node: on which node check/service is registered on
 - service_name: name of the service (this is the service name not the service ID)
+- check_id
 
 Fields:
-- check_id
 - check_name
 - service_id
 - status
+- passing
+- critical
+- warning
+
+`passing`, `critical`, and `warning` are integer representations of the health
+check state. A value of `1` represents that the status was the state of the
+the health check at this sample.
 
 ## Example output
 
 ```
 $ telegraf --config ./telegraf.conf -input-filter consul -test
 * Plugin: consul, Collection 1
-> consul_health_checks,host=wolfpit,node=consul-server-node check_id="serfHealth",check_name="Serf Health Status",service_id="",status="passing" 1464698464486439902
-> consul_health_checks,host=wolfpit,node=consul-server-node,service_name=www.example.com check_id="service:www-example-com.test01",check_name="Service 'www.example.com' check",service_id="www-example-com.test01",status="critical" 1464698464486519036
+> consul_health_checks,host=wolfpit,node=consul-server-node,check_id="serfHealth" check_name="Serf Health Status",service_id="",status="passing",passing=1i,critical=0i,warning=0i 1464698464486439902
+> consul_health_checks,host=wolfpit,node=consul-server-node,service_name=www.example.com,check_id="service:www-example-com.test01" check_name="Service 'www.example.com' check",service_id="www-example-com.test01",status="critical",passing=0i,critical=1i,warning=0i 1464698464486519036
 ```

plugins/inputs/consul/consul.go

@@ -95,13 +95,18 @@ func (c *Consul) GatherHealthCheck(acc telegraf.Accumulator, checks []*api.Healt
 		record := make(map[string]interface{})
 		tags := make(map[string]string)
 
-		record["check_id"] = check.CheckID
 		record["check_name"] = check.Name
 		record["service_id"] = check.ServiceID
+
 		record["status"] = check.Status
+		record["passing"] = 0
+		record["critical"] = 0
+		record["warning"] = 0
+		record[check.Status] = 1
 
 		tags["node"] = check.Node
 		tags["service_name"] = check.ServiceName
+		tags["check_id"] = check.CheckID
 
 		acc.AddFields("consul_health_checks", record, tags)
 	}

plugins/inputs/consul/consul_test.go

@@ -22,15 +22,18 @@ var sampleChecks = []*api.HealthCheck{
 
 func TestGatherHealtCheck(t *testing.T) {
 	expectedFields := map[string]interface{}{
-		"check_id":   "foo.health123",
 		"check_name": "foo.health",
 		"status":     "passing",
+		"passing":    1,
+		"critical":   0,
+		"warning":    0,
 		"service_id": "foo.123",
 	}
 
 	expectedTags := map[string]string{
 		"node":         "localhost",
 		"service_name": "foo",
+		"check_id":     "foo.health123",
 	}
 
 	var acc testutil.Accumulator

plugins/inputs/couchbase/couchbase.go

@@ -42,19 +42,17 @@ func (r *Couchbase) Gather(acc telegraf.Accumulator) error {
 
 	var wg sync.WaitGroup
 
-	var outerr error
-
 	for _, serv := range r.Servers {
 		wg.Add(1)
 		go func(serv string) {
 			defer wg.Done()
-			outerr = r.gatherServer(serv, acc, nil)
+			acc.AddError(r.gatherServer(serv, acc, nil))
 		}(serv)
 	}
 
 	wg.Wait()
 
-	return outerr
+	return nil
 }
 
 func (r *Couchbase) gatherServer(addr string, acc telegraf.Accumulator, pool *couchbase.Pool) error {

plugins/inputs/couchdb/couchdb.go

@@ -2,13 +2,11 @@ package couchdb
 
 import (
 	"encoding/json"
-	"errors"
 	"fmt"
 	"github.com/influxdata/telegraf"
 	"github.com/influxdata/telegraf/plugins/inputs"
 	"net/http"
 	"reflect"
-	"strings"
 	"sync"
 	"time"
 )
@@ -83,34 +81,20 @@ func (*CouchDB) SampleConfig() string {
 }
 
 func (c *CouchDB) Gather(accumulator telegraf.Accumulator) error {
-	errorChannel := make(chan error, len(c.HOSTs))
 	var wg sync.WaitGroup
 	for _, u := range c.HOSTs {
 		wg.Add(1)
 		go func(host string) {
 			defer wg.Done()
 			if err := c.fetchAndInsertData(accumulator, host); err != nil {
-				errorChannel <- fmt.Errorf("[host=%s]: %s", host, err)
+				accumulator.AddError(fmt.Errorf("[host=%s]: %s", host, err))
 			}
 		}(u)
 	}
 
 	wg.Wait()
-	close(errorChannel)
-
-	// If there weren't any errors, we can return nil now.
-	if len(errorChannel) == 0 {
-		return nil
-	}
-
-	// There were errors, so join them all together as one big error.
-	errorStrings := make([]string, 0, len(errorChannel))
-	for err := range errorChannel {
-		errorStrings = append(errorStrings, err.Error())
-	}
-
-	return errors.New(strings.Join(errorStrings, "\n"))
 
+	return nil
 }
 
 var tr = &http.Transport{

plugins/inputs/couchdb/couchdb_test.go

@@ -316,5 +316,5 @@ func TestBasic(t *testing.T) {
 	}
 
 	var acc testutil.Accumulator
-	require.NoError(t, plugin.Gather(&acc))
+	require.NoError(t, acc.GatherError(plugin.Gather))
 }

plugins/inputs/disque/disque.go

@@ -75,12 +75,11 @@ func (g *Disque) Gather(acc telegraf.Accumulator) error {
 
 	var wg sync.WaitGroup
 
-	var outerr error
-
 	for _, serv := range g.Servers {
 		u, err := url.Parse(serv)
 		if err != nil {
-			return fmt.Errorf("Unable to parse to address '%s': %s", serv, err)
+			acc.AddError(fmt.Errorf("Unable to parse to address '%s': %s", serv, err))
+			continue
 		} else if u.Scheme == "" {
 			// fallback to simple string based address (i.e. "10.0.0.1:10000")
 			u.Scheme = "tcp"
@@ -90,13 +89,13 @@ func (g *Disque) Gather(acc telegraf.Accumulator) error {
 		wg.Add(1)
 		go func(serv string) {
 			defer wg.Done()
-			outerr = g.gatherServer(u, acc)
+			acc.AddError(g.gatherServer(u, acc))
 		}(serv)
 	}
 
 	wg.Wait()
 
-	return outerr
+	return nil
 }
 
 const defaultPort = "7711"

plugins/inputs/disque/disque_test.go

@@ -51,7 +51,7 @@ func TestDisqueGeneratesMetrics(t *testing.T) {
 
 	var acc testutil.Accumulator
 
-	err = r.Gather(&acc)
+	err = acc.GatherError(r.Gather)
 	require.NoError(t, err)
 
 	fields := map[string]interface{}{
@@ -117,7 +117,7 @@ func TestDisqueCanPullStatsFromMultipleServers(t *testing.T) {
 
 	var acc testutil.Accumulator
 
-	err = r.Gather(&acc)
+	err = acc.GatherError(r.Gather)
 	require.NoError(t, err)
 
 	fields := map[string]interface{}{

plugins/inputs/dmcache/README.md

@@ -0,0 +1,47 @@
+# DMCache Input Plugin
+
+This plugin provide a native collection for dmsetup based statistics for dm-cache.
+
+This plugin requires sudo, that is why you should setup and be sure that the telegraf is able to execute sudo without a password.
+
+`sudo /sbin/dmsetup status --target cache` is the full command that telegraf will run for debugging purposes.
+
+### Configuration
+
+```toml
+[[inputs.dmcache]]
+  ## Whether to report per-device stats or not
+  per_device = true
+```
+
+### Measurements & Fields:
+
+- dmcache
+    - length
+    - target
+    - metadata_blocksize
+    - metadata_used
+    - metadata_total
+    - cache_blocksize
+    - cache_used
+    - cache_total
+    - read_hits
+    - read_misses
+    - write_hits
+    - write_misses
+    - demotions
+    - promotions
+    - dirty
+
+### Tags:
+
+- All measurements have the following tags:
+    - device
+
+### Example Output:
+
+```
+$ ./telegraf --test --config /etc/telegraf/telegraf.conf --input-filter dmcache
+* Plugin: inputs.dmcache, Collection 1
+> dmcache,device=example cache_blocksize=0i,read_hits=995134034411520i,read_misses=916807089127424i,write_hits=195107267543040i,metadata_used=12861440i,write_misses=563725346013184i,promotions=3265223720960i,dirty=0i,metadata_blocksize=0i,cache_used=1099511627776ii,cache_total=0i,length=0i,metadata_total=1073741824i,demotions=3265223720960i 1491482035000000000
+```

plugins/inputs/dmcache/dmcache.go

@@ -0,0 +1,33 @@
+package dmcache
+
+import (
+	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/plugins/inputs"
+)
+
+type DMCache struct {
+	PerDevice        bool `toml:"per_device"`
+	getCurrentStatus func() ([]string, error)
+}
+
+var sampleConfig = `
+  ## Whether to report per-device stats or not
+  per_device = true
+`
+
+func (c *DMCache) SampleConfig() string {
+	return sampleConfig
+}
+
+func (c *DMCache) Description() string {
+	return "Provide a native collection for dmsetup based statistics for dm-cache"
+}
+
+func init() {
+	inputs.Add("dmcache", func() telegraf.Input {
+		return &DMCache{
+			PerDevice:        true,
+			getCurrentStatus: dmSetupStatus,
+		}
+	})
+}

plugins/inputs/dmcache/dmcache_linux.go

@@ -0,0 +1,190 @@
+// +build linux
+
+package dmcache
+
+import (
+	"os/exec"
+	"strconv"
+	"strings"
+
+	"errors"
+
+	"github.com/influxdata/telegraf"
+)
+
+const metricName = "dmcache"
+
+type cacheStatus struct {
+	device            string
+	length            int
+	target            string
+	metadataBlocksize int
+	metadataUsed      int
+	metadataTotal     int
+	cacheBlocksize    int
+	cacheUsed         int
+	cacheTotal        int
+	readHits          int
+	readMisses        int
+	writeHits         int
+	writeMisses       int
+	demotions         int
+	promotions        int
+	dirty             int
+}
+
+func (c *DMCache) Gather(acc telegraf.Accumulator) error {
+	outputLines, err := c.getCurrentStatus()
+	if err != nil {
+		return err
+	}
+
+	totalStatus := cacheStatus{}
+
+	for _, s := range outputLines {
+		status, err := parseDMSetupStatus(s)
+		if err != nil {
+			return err
+		}
+
+		if c.PerDevice {
+			tags := map[string]string{"device": status.device}
+			acc.AddFields(metricName, toFields(status), tags)
+		}
+		aggregateStats(&totalStatus, status)
+	}
+
+	acc.AddFields(metricName, toFields(totalStatus), map[string]string{"device": "all"})
+
+	return nil
+}
+
+func parseDMSetupStatus(line string) (cacheStatus, error) {
+	var err error
+	parseError := errors.New("Output from dmsetup could not be parsed")
+	status := cacheStatus{}
+	values := strings.Fields(line)
+	if len(values) < 15 {
+		return cacheStatus{}, parseError
+	}
+
+	status.device = strings.TrimRight(values[0], ":")
+	status.length, err = strconv.Atoi(values[2])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	status.target = values[3]
+	status.metadataBlocksize, err = strconv.Atoi(values[4])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	metadata := strings.Split(values[5], "/")
+	if len(metadata) != 2 {
+		return cacheStatus{}, parseError
+	}
+	status.metadataUsed, err = strconv.Atoi(metadata[0])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	status.metadataTotal, err = strconv.Atoi(metadata[1])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	status.cacheBlocksize, err = strconv.Atoi(values[6])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	cache := strings.Split(values[7], "/")
+	if len(cache) != 2 {
+		return cacheStatus{}, parseError
+	}
+	status.cacheUsed, err = strconv.Atoi(cache[0])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	status.cacheTotal, err = strconv.Atoi(cache[1])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	status.readHits, err = strconv.Atoi(values[8])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	status.readMisses, err = strconv.Atoi(values[9])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	status.writeHits, err = strconv.Atoi(values[10])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	status.writeMisses, err = strconv.Atoi(values[11])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	status.demotions, err = strconv.Atoi(values[12])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	status.promotions, err = strconv.Atoi(values[13])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+	status.dirty, err = strconv.Atoi(values[14])
+	if err != nil {
+		return cacheStatus{}, err
+	}
+
+	return status, nil
+}
+
+func aggregateStats(totalStatus *cacheStatus, status cacheStatus) {
+	totalStatus.length += status.length
+	totalStatus.metadataBlocksize += status.metadataBlocksize
+	totalStatus.metadataUsed += status.metadataUsed
+	totalStatus.metadataTotal += status.metadataTotal
+	totalStatus.cacheBlocksize += status.cacheBlocksize
+	totalStatus.cacheUsed += status.cacheUsed
+	totalStatus.cacheTotal += status.cacheTotal
+	totalStatus.readHits += status.readHits
+	totalStatus.readMisses += status.readMisses
+	totalStatus.writeHits += status.writeHits
+	totalStatus.writeMisses += status.writeMisses
+	totalStatus.demotions += status.demotions
+	totalStatus.promotions += status.promotions
+	totalStatus.dirty += status.dirty
+}
+
+func toFields(status cacheStatus) map[string]interface{} {
+	fields := make(map[string]interface{})
+	fields["length"] = status.length
+	fields["metadata_blocksize"] = status.metadataBlocksize
+	fields["metadata_used"] = status.metadataUsed
+	fields["metadata_total"] = status.metadataTotal
+	fields["cache_blocksize"] = status.cacheBlocksize
+	fields["cache_used"] = status.cacheUsed
+	fields["cache_total"] = status.cacheTotal
+	fields["read_hits"] = status.readHits
+	fields["read_misses"] = status.readMisses
+	fields["write_hits"] = status.writeHits
+	fields["write_misses"] = status.writeMisses
+	fields["demotions"] = status.demotions
+	fields["promotions"] = status.promotions
+	fields["dirty"] = status.dirty
+	return fields
+}
+
+func dmSetupStatus() ([]string, error) {
+	out, err := exec.Command("/bin/sh", "-c", "sudo /sbin/dmsetup status --target cache").Output()
+	if err != nil {
+		return nil, err
+	}
+	if string(out) == "No devices found\n" {
+		return []string{}, nil
+	}
+
+	outString := strings.TrimRight(string(out), "\n")
+	status := strings.Split(outString, "\n")
+
+	return status, nil
+}

plugins/inputs/dmcache/dmcache_notlinux.go

@@ -0,0 +1,15 @@
+// +build !linux
+
+package dmcache
+
+import (
+	"github.com/influxdata/telegraf"
+)
+
+func (c *DMCache) Gather(acc telegraf.Accumulator) error {
+	return nil
+}
+
+func dmSetupStatus() ([]string, error) {
+	return []string{}, nil
+}

plugins/inputs/dmcache/dmcache_test.go

@@ -0,0 +1,169 @@
+package dmcache
+
+import (
+	"errors"
+	"testing"
+
+	"github.com/influxdata/telegraf/testutil"
+	"github.com/stretchr/testify/require"
+)
+
+var (
+	measurement              = "dmcache"
+	badFormatOutput          = []string{"cs-1: 0 4883791872 cache 8 1018/1501122 512 7/464962 139 352643 "}
+	good2DevicesFormatOutput = []string{
+		"cs-1: 0 4883791872 cache 8 1018/1501122 512 7/464962 139 352643 15 46 0 7 0 1 writeback 2 migration_threshold 2048 mq 10 random_threshold 4 sequential_threshold 512 discard_promote_adjustment 1 read_promote_adjustment 4 write_promote_adjustment 8",
+		"cs-2: 0 4294967296 cache 8 72352/1310720 128 26/24327168 2409 286 265 524682 0 0 0 1 writethrough 2 migration_threshold 2048 mq 10 random_threshold 4 sequential_threshold 512 discard_promote_adjustment 1 read_promote_adjustment 4 write_promote_adjustment 8",
+	}
+)
+
+func TestPerDeviceGoodOutput(t *testing.T) {
+	var acc testutil.Accumulator
+	var plugin = &DMCache{
+		PerDevice: true,
+		getCurrentStatus: func() ([]string, error) {
+			return good2DevicesFormatOutput, nil
+		},
+	}
+
+	err := plugin.Gather(&acc)
+	require.NoError(t, err)
+
+	tags1 := map[string]string{
+		"device": "cs-1",
+	}
+	fields1 := map[string]interface{}{
+		"length":             4883791872,
+		"metadata_blocksize": 8,
+		"metadata_used":      1018,
+		"metadata_total":     1501122,
+		"cache_blocksize":    512,
+		"cache_used":         7,
+		"cache_total":        464962,
+		"read_hits":          139,
+		"read_misses":        352643,
+		"write_hits":         15,
+		"write_misses":       46,
+		"demotions":          0,
+		"promotions":         7,
+		"dirty":              0,
+	}
+	acc.AssertContainsTaggedFields(t, measurement, fields1, tags1)
+
+	tags2 := map[string]string{
+		"device": "cs-2",
+	}
+	fields2 := map[string]interface{}{
+		"length":             4294967296,
+		"metadata_blocksize": 8,
+		"metadata_used":      72352,
+		"metadata_total":     1310720,
+		"cache_blocksize":    128,
+		"cache_used":         26,
+		"cache_total":        24327168,
+		"read_hits":          2409,
+		"read_misses":        286,
+		"write_hits":         265,
+		"write_misses":       524682,
+		"demotions":          0,
+		"promotions":         0,
+		"dirty":              0,
+	}
+	acc.AssertContainsTaggedFields(t, measurement, fields2, tags2)
+
+	tags3 := map[string]string{
+		"device": "all",
+	}
+
+	fields3 := map[string]interface{}{
+		"length":             9178759168,
+		"metadata_blocksize": 16,
+		"metadata_used":      73370,
+		"metadata_total":     2811842,
+		"cache_blocksize":    640,
+		"cache_used":         33,
+		"cache_total":        24792130,
+		"read_hits":          2548,
+		"read_misses":        352929,
+		"write_hits":         280,
+		"write_misses":       524728,
+		"demotions":          0,
+		"promotions":         7,
+		"dirty":              0,
+	}
+	acc.AssertContainsTaggedFields(t, measurement, fields3, tags3)
+}
+
+func TestNotPerDeviceGoodOutput(t *testing.T) {
+	var acc testutil.Accumulator
+	var plugin = &DMCache{
+		PerDevice: false,
+		getCurrentStatus: func() ([]string, error) {
+			return good2DevicesFormatOutput, nil
+		},
+	}
+
+	err := plugin.Gather(&acc)
+	require.NoError(t, err)
+
+	tags := map[string]string{
+		"device": "all",
+	}
+
+	fields := map[string]interface{}{
+		"length":             9178759168,
+		"metadata_blocksize": 16,
+		"metadata_used":      73370,
+		"metadata_total":     2811842,
+		"cache_blocksize":    640,
+		"cache_used":         33,
+		"cache_total":        24792130,
+		"read_hits":          2548,
+		"read_misses":        352929,
+		"write_hits":         280,
+		"write_misses":       524728,
+		"demotions":          0,
+		"promotions":         7,
+		"dirty":              0,
+	}
+	acc.AssertContainsTaggedFields(t, measurement, fields, tags)
+}
+
+func TestNoDevicesOutput(t *testing.T) {
+	var acc testutil.Accumulator
+	var plugin = &DMCache{
+		PerDevice: true,
+		getCurrentStatus: func() ([]string, error) {
+			return []string{}, nil
+		},
+	}
+
+	err := plugin.Gather(&acc)
+	require.NoError(t, err)
+}
+
+func TestErrorDuringGettingStatus(t *testing.T) {
+	var acc testutil.Accumulator
+	var plugin = &DMCache{
+		PerDevice: true,
+		getCurrentStatus: func() ([]string, error) {
+			return nil, errors.New("dmsetup doesn't exist")
+		},
+	}
+
+	err := plugin.Gather(&acc)
+	require.Error(t, err)
+}
+
+func TestBadFormatOfStatus(t *testing.T) {
+	var acc testutil.Accumulator
+	var plugin = &DMCache{
+		PerDevice: true,
+		getCurrentStatus: func() ([]string, error) {
+			return badFormatOutput, nil
+		},
+	}
+
+	err := plugin.Gather(&acc)
+	require.Error(t, err)
+}

plugins/inputs/dns_query/dns_query.go

@@ -9,7 +9,6 @@ import (
 	"time"
 
 	"github.com/influxdata/telegraf"
-	"github.com/influxdata/telegraf/internal/errchan"
 	"github.com/influxdata/telegraf/plugins/inputs"
 )
 
@@ -58,11 +57,10 @@ func (d *DnsQuery) Description() string {
 func (d *DnsQuery) Gather(acc telegraf.Accumulator) error {
 	d.setDefaultValues()
 
-	errChan := errchan.New(len(d.Domains) * len(d.Servers))
 	for _, domain := range d.Domains {
 		for _, server := range d.Servers {
 			dnsQueryTime, err := d.getDnsQueryTime(domain, server)
-			errChan.C <- err
+			acc.AddError(err)
 			tags := map[string]string{
 				"server":      server,
 				"domain":      domain,
@@ -74,7 +72,7 @@ func (d *DnsQuery) Gather(acc telegraf.Accumulator) error {
 		}
 	}
 
-	return errChan.Error()
+	return nil
 }
 
 func (d *DnsQuery) setDefaultValues() {

plugins/inputs/dns_query/dns_query_test.go

@@ -24,7 +24,7 @@ func TestGathering(t *testing.T) {
 	}
 	var acc testutil.Accumulator
 
-	err := dnsConfig.Gather(&acc)
+	err := acc.GatherError(dnsConfig.Gather)
 	assert.NoError(t, err)
 	metric, ok := acc.Get("dns_query")
 	require.True(t, ok)
@@ -44,7 +44,7 @@ func TestGatheringMxRecord(t *testing.T) {
 	var acc testutil.Accumulator
 	dnsConfig.RecordType = "MX"
 
-	err := dnsConfig.Gather(&acc)
+	err := acc.GatherError(dnsConfig.Gather)
 	assert.NoError(t, err)
 	metric, ok := acc.Get("dns_query")
 	require.True(t, ok)
@@ -70,7 +70,7 @@ func TestGatheringRootDomain(t *testing.T) {
 	}
 	fields := map[string]interface{}{}
 
-	err := dnsConfig.Gather(&acc)
+	err := acc.GatherError(dnsConfig.Gather)
 	assert.NoError(t, err)
 	metric, ok := acc.Get("dns_query")
 	require.True(t, ok)
@@ -96,7 +96,7 @@ func TestMetricContainsServerAndDomainAndRecordTypeTags(t *testing.T) {
 	}
 	fields := map[string]interface{}{}
 
-	err := dnsConfig.Gather(&acc)
+	err := acc.GatherError(dnsConfig.Gather)
 	assert.NoError(t, err)
 	metric, ok := acc.Get("dns_query")
 	require.True(t, ok)
@@ -121,7 +121,7 @@ func TestGatheringTimeout(t *testing.T) {
 
 	channel := make(chan error, 1)
 	go func() {
-		channel <- dnsConfig.Gather(&acc)
+		channel <- acc.GatherError(dnsConfig.Gather)
 	}()
 	select {
 	case res := <-channel:

plugins/inputs/docker/README.md

@@ -16,12 +16,26 @@ for the stat structure can be found
 ```
 # Read metrics about docker containers
 [[inputs.docker]]
-  # Docker Endpoint
-  #   To use TCP, set endpoint = "tcp://[ip]:[port]"
-  #   To use environment variables (ie, docker-machine), set endpoint = "ENV"
+  ## Docker Endpoint
+  ##   To use TCP, set endpoint = "tcp://[ip]:[port]"
+  ##   To use environment variables (ie, docker-machine), set endpoint = "ENV"
   endpoint = "unix:///var/run/docker.sock"
-  # Only collect metrics for these containers, collect all if empty
+  ## Only collect metrics for these containers, collect all if empty
   container_names = []
+  ## Timeout for docker list, info, and stats commands
+  timeout = "5s"
+
+  ## Whether to report for each container per-device blkio (8:0, 8:1...) and
+  ## network (eth0, eth1, ...) stats or not
+  perdevice = true
+  ## Whether to report for each container total blkio and network stats or not
+  total = false
+  
+  ## docker labels to include and exclude as tags.  Globs accepted.
+  ## Note that an empty array for both will include all labels as tags
+  docker_label_include = []
+  docker_label_exclude = []
+  
 ```
 
 ### Measurements & Fields:
@@ -122,30 +136,32 @@ based on the availability of per-cpu stats on your system.
 
 
 ### Tags:
-
+#### Docker Engine tags
 - docker (memory_total)
     - unit=bytes
+    - engine_host
 - docker (pool_blocksize)
     - unit=bytes
+    - engine_host
 - docker_data
     - unit=bytes
+    - engine_host
 - docker_metadata
     - unit=bytes
+    - engine_host
 
+#### Docker Container tags
+- Tags on all containers:
+    - engine_host
+    - container_image
+    - container_name
+    - container_version
 - docker_container_mem specific:
-    - container_image
-    - container_name
 - docker_container_cpu specific:
-    - container_image
-    - container_name
     - cpu
 - docker_container_net specific:
-    - container_image
-    - container_name
     - network
 - docker_container_blkio specific:
-    - container_image
-    - container_name
     - device
 
 ### Example Output:

plugins/inputs/docker/docker.go

@@ -1,42 +1,82 @@
-package system
+package docker
 
 import (
+	"context"
 	"encoding/json"
 	"fmt"
 	"io"
-	"log"
 	"regexp"
 	"strconv"
 	"strings"
 	"sync"
 	"time"
 
-	"golang.org/x/net/context"
-
-	"github.com/docker/engine-api/client"
-	"github.com/docker/engine-api/types"
+	"github.com/docker/docker/api/types"
+	"github.com/docker/docker/client"
 	"github.com/influxdata/telegraf"
+	"github.com/influxdata/telegraf/filter"
 	"github.com/influxdata/telegraf/internal"
 	"github.com/influxdata/telegraf/plugins/inputs"
 )
 
+type DockerLabelFilter struct {
+	labelInclude filter.Filter
+	labelExclude filter.Filter
+}
+
 // Docker object
 type Docker struct {
 	Endpoint       string
 	ContainerNames []string
 	Timeout        internal.Duration
-	PerDevice      bool `toml:"perdevice"`
-	Total          bool `toml:"total"`
+	PerDevice      bool     `toml:"perdevice"`
+	Total          bool     `toml:"total"`
+	LabelInclude   []string `toml:"docker_label_include"`
+	LabelExclude   []string `toml:"docker_label_exclude"`
 
-	client      DockerClient
+	LabelFilter DockerLabelFilter
+
+	client      *client.Client
 	engine_host string
+
+	testing             bool
+	labelFiltersCreated bool
 }
 
-// DockerClient interface, useful for testing
-type DockerClient interface {
-	Info(ctx context.Context) (types.Info, error)
-	ContainerList(ctx context.Context, options types.ContainerListOptions) ([]types.Container, error)
-	ContainerStats(ctx context.Context, containerID string, stream bool) (io.ReadCloser, error)
+// infoWrapper wraps client.Client.List for testing.
+func infoWrapper(c *client.Client, ctx context.Context) (types.Info, error) {
+	if c != nil {
+		return c.Info(ctx)
+	}
+	fc := FakeDockerClient{}
+	return fc.Info(ctx)
+}
+
+// listWrapper wraps client.Client.ContainerList for testing.
+func listWrapper(
+	c *client.Client,
+	ctx context.Context,
+	options types.ContainerListOptions,
+) ([]types.Container, error) {
+	if c != nil {
+		return c.ContainerList(ctx, options)
+	}
+	fc := FakeDockerClient{}
+	return fc.ContainerList(ctx, options)
+}
+
+// statsWrapper wraps client.Client.ContainerStats for testing.
+func statsWrapper(
+	c *client.Client,
+	ctx context.Context,
+	containerID string,
+	stream bool,
+) (types.ContainerStats, error) {
+	if c != nil {
+		return c.ContainerStats(ctx, containerID, stream)
+	}
+	fc := FakeDockerClient{}
+	return fc.ContainerStats(ctx, containerID, stream)
 }
 
 // KB, MB, GB, TB, PB...human friendly
@@ -68,6 +108,10 @@ var sampleConfig = `
   ## Whether to report for each container total blkio and network stats or not
   total = false
 
+  ## docker labels to include and exclude as tags.  Globs accepted.
+  ## Note that an empty array for both will include all labels as tags
+  docker_label_include = []
+  docker_label_exclude = []
 `
 
 // Description returns input description
@@ -80,7 +124,7 @@ func (d *Docker) SampleConfig() string { return sampleConfig }
 
 // Gather starts stats collection
 func (d *Docker) Gather(acc telegraf.Accumulator) error {
-	if d.client == nil {
+	if d.client == nil && !d.testing {
 		var c *client.Client
 		var err error
 		defaultHeaders := map[string]string{"User-Agent": "engine-api-cli-1.0"}
@@ -102,18 +146,26 @@ func (d *Docker) Gather(acc telegraf.Accumulator) error {
 		}
 		d.client = c
 	}
+	// Create label filters if not already created
+	if !d.labelFiltersCreated {
+		err := d.createLabelFilters()
+		if err != nil {
+			return err
+		}
+		d.labelFiltersCreated = true
+	}
 
 	// Get daemon info
 	err := d.gatherInfo(acc)
 	if err != nil {
-		fmt.Println(err.Error())
+		acc.AddError(err)
 	}
 
 	// List containers
 	opts := types.ContainerListOptions{}
 	ctx, cancel := context.WithTimeout(context.Background(), d.Timeout.Duration)
 	defer cancel()
-	containers, err := d.client.ContainerList(ctx, opts)
+	containers, err := listWrapper(d.client, ctx, opts)
 	if err != nil {
 		return err
 	}
@@ -126,8 +178,8 @@ func (d *Docker) Gather(acc telegraf.Accumulator) error {
 			defer wg.Done()
 			err := d.gatherContainer(c, acc)
 			if err != nil {
-				log.Printf("E! Error gathering container %s stats: %s\n",
-					c.Names, err.Error())
+				acc.AddError(fmt.Errorf("E! Error gathering container %s stats: %s\n",
+					c.Names, err.Error()))
 			}
 		}(container)
 	}
@@ -144,7 +196,7 @@ func (d *Docker) gatherInfo(acc telegraf.Accumulator) error {
 	// Get info from docker daemon
 	ctx, cancel := context.WithTimeout(context.Background(), d.Timeout.Duration)
 	defer cancel()
-	info, err := d.client.Info(ctx)
+	info, err := infoWrapper(d.client, ctx)
 	if err != nil {
 		return err
 	}
@@ -221,14 +273,18 @@ func (d *Docker) gatherContainer(
 		cname = strings.TrimPrefix(container.Names[0], "/")
 	}
 
-	// the image name sometimes has a version part.
-	//   ie, rabbitmq:3-management
-	imageParts := strings.Split(container.Image, ":")
-	imageName := imageParts[0]
+	// the image name sometimes has a version part, or a private repo
+	//   ie, rabbitmq:3-management or docker.someco.net:4443/rabbitmq:3-management
+	imageName := ""
 	imageVersion := "unknown"
-	if len(imageParts) > 1 {
-		imageVersion = imageParts[1]
+	i := strings.LastIndex(container.Image, ":") // index of last ':' character
+	if i > -1 {
+		imageVersion = container.Image[i+1:]
+		imageName = container.Image[:i]
+	} else {
+		imageName = container.Image
 	}
+
 	tags := map[string]string{
 		"engine_host":       d.engine_host,
 		"container_name":    cname,
@@ -243,12 +299,12 @@ func (d *Docker) gatherContainer(
 
 	ctx, cancel := context.WithTimeout(context.Background(), d.Timeout.Duration)
 	defer cancel()
-	r, err := d.client.ContainerStats(ctx, container.ID, false)
+	r, err := statsWrapper(d.client, ctx, container.ID, false)
 	if err != nil {
 		return fmt.Errorf("Error getting docker stats: %s", err.Error())
 	}
-	defer r.Close()
-	dec := json.NewDecoder(r)
+	defer r.Body.Close()
+	dec := json.NewDecoder(r.Body)
 	if err = dec.Decode(&v); err != nil {
 		if err == io.EOF {
 			return nil
@@ -258,7 +314,11 @@ func (d *Docker) gatherContainer(
 
 	// Add labels to tags
 	for k, label := range container.Labels {
-		tags[k] = label
+		if len(d.LabelInclude) == 0 || d.LabelFilter.labelInclude.Match(k) {
+			if len(d.LabelExclude) == 0 || !d.LabelFilter.labelExclude.Match(k) {
+				tags[k] = label
+			}
+		}
 	}
 
 	gatherContainerStats(v, acc, tags, container.ID, d.PerDevice, d.Total)
@@ -364,11 +424,22 @@ func gatherContainerStats(
 				if field == "container_id" {
 					continue
 				}
+
+				var uintV uint64
+				switch v := value.(type) {
+				case uint64:
+					uintV = v
+				case int64:
+					uintV = uint64(v)
+				default:
+					continue
+				}
+
 				_, ok := totalNetworkStatMap[field]
 				if ok {
-					totalNetworkStatMap[field] = totalNetworkStatMap[field].(uint64) + value.(uint64)
+					totalNetworkStatMap[field] = totalNetworkStatMap[field].(uint64) + uintV
 				} else {
-					totalNetworkStatMap[field] = value
+					totalNetworkStatMap[field] = uintV
 				}
 			}
 		}
@@ -487,11 +558,22 @@ func gatherBlockIOMetrics(
 				if field == "container_id" {
 					continue
 				}
+
+				var uintV uint64
+				switch v := value.(type) {
+				case uint64:
+					uintV = v
+				case int64:
+					uintV = uint64(v)
+				default:
+					continue
+				}
+
 				_, ok := totalStatMap[field]
 				if ok {
-					totalStatMap[field] = totalStatMap[field].(uint64) + value.(uint64)
+					totalStatMap[field] = totalStatMap[field].(uint64) + uintV
 				} else {
-					totalStatMap[field] = value
+					totalStatMap[field] = uintV
 				}
 			}
 		}
@@ -542,11 +624,32 @@ func parseSize(sizeStr string) (int64, error) {
 	return int64(size), nil
 }
 
+func (d *Docker) createLabelFilters() error {
+	if len(d.LabelInclude) != 0 && d.LabelFilter.labelInclude == nil {
+		var err error
+		d.LabelFilter.labelInclude, err = filter.Compile(d.LabelInclude)
+		if err != nil {
+			return err
+		}
+	}
+
+	if len(d.LabelExclude) != 0 && d.LabelFilter.labelExclude == nil {
+		var err error
+		d.LabelFilter.labelExclude, err = filter.Compile(d.LabelExclude)
+		if err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
 func init() {
 	inputs.Add("docker", func() telegraf.Input {
 		return &Docker{
-			PerDevice: true,
-			Timeout:   internal.Duration{Duration: time.Second * 5},
+			PerDevice:           true,
+			Timeout:             internal.Duration{Duration: time.Second * 5},
+			labelFiltersCreated: false,
 		}
 	})
 }

plugins/inputs/docker/docker_test.go

@@ -1,18 +1,12 @@
-package system
+package docker
 
 import (
-	"io"
-	"io/ioutil"
-	"strings"
 	"testing"
 	"time"
 
-	"golang.org/x/net/context"
-
-	"github.com/docker/engine-api/types"
-	"github.com/docker/engine-api/types/registry"
 	"github.com/influxdata/telegraf/testutil"
 
+	"github.com/docker/docker/api/types"
 	"github.com/stretchr/testify/require"
 )
 
@@ -250,147 +244,65 @@ func testStats() *types.StatsJSON {
 	return stats
 }
 
-type FakeDockerClient struct {
+var gatherLabelsTests = []struct {
+	include     []string
+	exclude     []string
+	expected    []string
+	notexpected []string
+}{
+	{[]string{}, []string{}, []string{"label1", "label2"}, []string{}},
+	{[]string{"*"}, []string{}, []string{"label1", "label2"}, []string{}},
+	{[]string{"lab*"}, []string{}, []string{"label1", "label2"}, []string{}},
+	{[]string{"label1"}, []string{}, []string{"label1"}, []string{"label2"}},
+	{[]string{"label1*"}, []string{}, []string{"label1"}, []string{"label2"}},
+	{[]string{}, []string{"*"}, []string{}, []string{"label1", "label2"}},
+	{[]string{}, []string{"lab*"}, []string{}, []string{"label1", "label2"}},
+	{[]string{}, []string{"label1"}, []string{"label2"}, []string{"label1"}},
+	{[]string{"*"}, []string{"*"}, []string{}, []string{"label1", "label2"}},
 }
 
-func (d FakeDockerClient) Info(ctx context.Context) (types.Info, error) {
-	env := types.Info{
-		Containers:         108,
-		ContainersRunning:  98,
-		ContainersStopped:  6,
-		ContainersPaused:   3,
-		OomKillDisable:     false,
-		SystemTime:         "2016-02-24T00:55:09.15073105-05:00",
-		NEventsListener:    0,
-		ID:                 "5WQQ:TFWR:FDNG:OKQ3:37Y4:FJWG:QIKK:623T:R3ME:QTKB:A7F7:OLHD",
-		Debug:              false,
-		LoggingDriver:      "json-file",
-		KernelVersion:      "4.3.0-1-amd64",
-		IndexServerAddress: "https://index.docker.io/v1/",
-		MemTotal:           3840757760,
-		Images:             199,
-		CPUCfsQuota:        true,
-		Name:               "absol",
-		SwapLimit:          false,
-		IPv4Forwarding:     true,
-		ExecutionDriver:    "native-0.2",
-		ExperimentalBuild:  false,
-		CPUCfsPeriod:       true,
-		RegistryConfig: &registry.ServiceConfig{
-			IndexConfigs: map[string]*registry.IndexInfo{
-				"docker.io": {
-					Name:     "docker.io",
-					Mirrors:  []string{},
-					Official: true,
-					Secure:   true,
-				},
-			}, InsecureRegistryCIDRs: []*registry.NetIPNet{{IP: []byte{127, 0, 0, 0}, Mask: []byte{255, 0, 0, 0}}}, Mirrors: []string{}},
-		OperatingSystem:   "Linux Mint LMDE (containerized)",
-		BridgeNfIptables:  true,
-		HTTPSProxy:        "",
-		Labels:            []string{},
-		MemoryLimit:       false,
-		DriverStatus:      [][2]string{{"Pool Name", "docker-8:1-1182287-pool"}, {"Pool Blocksize", "65.54 kB"}, {"Backing Filesystem", "extfs"}, {"Data file", "/dev/loop0"}, {"Metadata file", "/dev/loop1"}, {"Data Space Used", "17.3 GB"}, {"Data Space Total", "107.4 GB"}, {"Data Space Available", "36.53 GB"}, {"Metadata Space Used", "20.97 MB"}, {"Metadata Space Total", "2.147 GB"}, {"Metadata Space Available", "2.127 GB"}, {"Udev Sync Supported", "true"}, {"Deferred Removal Enabled", "false"}, {"Data loop file", "/var/lib/docker/devicemapper/devicemapper/data"}, {"Metadata loop file", "/var/lib/docker/devicemapper/devicemapper/metadata"}, {"Library Version", "1.02.115 (2016-01-25)"}},
-		NFd:               19,
-		HTTPProxy:         "",
-		Driver:            "devicemapper",
-		NGoroutines:       39,
-		NCPU:              4,
-		DockerRootDir:     "/var/lib/docker",
-		NoProxy:           "",
-		BridgeNfIP6tables: true,
+func TestDockerGatherLabels(t *testing.T) {
+	for _, tt := range gatherLabelsTests {
+		var acc testutil.Accumulator
+		d := Docker{
+			client:  nil,
+			testing: true,
+		}
+
+		for _, label := range tt.include {
+			d.LabelInclude = append(d.LabelInclude, label)
+		}
+		for _, label := range tt.exclude {
+			d.LabelExclude = append(d.LabelExclude, label)
+		}
+
+		err := d.Gather(&acc)
+		require.NoError(t, err)
+
+		for _, label := range tt.expected {
+			if !acc.HasTag("docker_container_cpu", label) {
+				t.Errorf("Didn't get expected label of %s.  Test was:  Include: %s  Exclude %s",
+					label, tt.include, tt.exclude)
+			}
+		}
+
+		for _, label := range tt.notexpected {
+			if acc.HasTag("docker_container_cpu", label) {
+				t.Errorf("Got unexpected label of %s.  Test was:  Include: %s  Exclude %s",
+					label, tt.include, tt.exclude)
+			}
+		}
 	}
-	return env, nil
-}
-
-func (d FakeDockerClient) ContainerList(octx context.Context, options types.ContainerListOptions) ([]types.Container, error) {
-	container1 := types.Container{
-		ID:      "e2173b9478a6ae55e237d4d74f8bbb753f0817192b5081334dc78476296b7dfb",
-		Names:   []string{"/etcd"},
-		Image:   "quay.io/coreos/etcd:v2.2.2",
-		Command: "/etcd -name etcd0 -advertise-client-urls http://localhost:2379 -listen-client-urls http://0.0.0.0:2379",
-		Created: 1455941930,
-		Status:  "Up 4 hours",
-		Ports: []types.Port{
-			types.Port{
-				PrivatePort: 7001,
-				PublicPort:  0,
-				Type:        "tcp",
-			},
-			types.Port{
-				PrivatePort: 4001,
-				PublicPort:  0,
-				Type:        "tcp",
-			},
-			types.Port{
-				PrivatePort: 2380,
-				PublicPort:  0,
-				Type:        "tcp",
-			},
-			types.Port{
-				PrivatePort: 2379,
-				PublicPort:  2379,
-				Type:        "tcp",
-				IP:          "0.0.0.0",
-			},
-		},
-		SizeRw:     0,
-		SizeRootFs: 0,
-	}
-	container2 := types.Container{
-		ID:      "b7dfbb9478a6ae55e237d4d74f8bbb753f0817192b5081334dc78476296e2173",
-		Names:   []string{"/etcd2"},
-		Image:   "quay.io/coreos/etcd:v2.2.2",
-		Command: "/etcd -name etcd2 -advertise-client-urls http://localhost:2379 -listen-client-urls http://0.0.0.0:2379",
-		Created: 1455941933,
-		Status:  "Up 4 hours",
-		Ports: []types.Port{
-			types.Port{
-				PrivatePort: 7002,
-				PublicPort:  0,
-				Type:        "tcp",
-			},
-			types.Port{
-				PrivatePort: 4002,
-				PublicPort:  0,
-				Type:        "tcp",
-			},
-			types.Port{
-				PrivatePort: 2381,
-				PublicPort:  0,
-				Type:        "tcp",
-			},
-			types.Port{
-				PrivatePort: 2382,
-				PublicPort:  2382,
-				Type:        "tcp",
-				IP:          "0.0.0.0",
-			},
-		},
-		SizeRw:     0,
-		SizeRootFs: 0,
-	}
-
-	containers := []types.Container{container1, container2}
-	return containers, nil
-
-	//#{e6a96c84ca91a5258b7cb752579fb68826b68b49ff957487695cd4d13c343b44 titilambert/snmpsim /bin/sh -c 'snmpsimd --agent-udpv4-endpoint=0.0.0.0:31161 --process-user=root --process-group=user' 1455724831 Up 4 hours [{31161 31161 udp 0.0.0.0}] 0 0 [/snmp] map[]}]2016/02/24 01:05:01 Gathered metrics, (3s interval), from 1 inputs in 1.233836656s
-}
-
-func (d FakeDockerClient) ContainerStats(ctx context.Context, containerID string, stream bool) (io.ReadCloser, error) {
-	var stat io.ReadCloser
-	jsonStat := `{"read":"2016-02-24T11:42:27.472459608-05:00","memory_stats":{"stats":{},"limit":18935443456},"blkio_stats":{"io_service_bytes_recursive":[{"major":252,"minor":1,"op":"Read","value":753664},{"major":252,"minor":1,"op":"Write"},{"major":252,"minor":1,"op":"Sync"},{"major":252,"minor":1,"op":"Async","value":753664},{"major":252,"minor":1,"op":"Total","value":753664}],"io_serviced_recursive":[{"major":252,"minor":1,"op":"Read","value":26},{"major":252,"minor":1,"op":"Write"},{"major":252,"minor":1,"op":"Sync"},{"major":252,"minor":1,"op":"Async","value":26},{"major":252,"minor":1,"op":"Total","value":26}]},"cpu_stats":{"cpu_usage":{"percpu_usage":[17871,4959158,1646137,1231652,11829401,244656,369972,0],"usage_in_usermode":10000000,"total_usage":20298847},"system_cpu_usage":24052607520000000,"throttling_data":{}},"precpu_stats":{"cpu_usage":{"percpu_usage":[17871,4959158,1646137,1231652,11829401,244656,369972,0],"usage_in_usermode":10000000,"total_usage":20298847},"system_cpu_usage":24052599550000000,"throttling_data":{}}}`
-	stat = ioutil.NopCloser(strings.NewReader(jsonStat))
-	return stat, nil
 }
 
 func TestDockerGatherInfo(t *testing.T) {
 	var acc testutil.Accumulator
-	client := FakeDockerClient{}
-	d := Docker{client: client}
-
-	err := d.Gather(&acc)
+	d := Docker{
+		client:  nil,
+		testing: true,
+	}
 
+	err := acc.GatherError(d.Gather)
 	require.NoError(t, err)
 
 	acc.AssertContainsTaggedFields(t,
@@ -429,10 +341,12 @@ func TestDockerGatherInfo(t *testing.T) {
 		},
 		map[string]string{
 			"container_name":    "etcd2",
-			"container_image":   "quay.io/coreos/etcd",
+			"container_image":   "quay.io:4443/coreos/etcd",
 			"cpu":               "cpu3",
 			"container_version": "v2.2.2",
 			"engine_host":       "absol",
+			"label1":            "test_value_1",
+			"label2":            "test_value_2",
 		},
 	)
 	acc.AssertContainsTaggedFields(t,
@@ -477,8 +391,10 @@ func TestDockerGatherInfo(t *testing.T) {
 		map[string]string{
 			"engine_host":       "absol",
 			"container_name":    "etcd2",
-			"container_image":   "quay.io/coreos/etcd",
+			"container_image":   "quay.io:4443/coreos/etcd",
 			"container_version": "v2.2.2",
+			"label1":            "test_value_1",
+			"label2":            "test_value_2",
 		},
 	)
 

plugins/inputs/docker/fake_client.go

@@ -0,0 +1,151 @@
+package docker
+
+import (
+	"context"
+	"io/ioutil"
+	"strings"
+
+	"github.com/docker/docker/api/types"
+	"github.com/docker/docker/api/types/registry"
+)
+
+type FakeDockerClient struct {
+}
+
+func (d FakeDockerClient) Info(ctx context.Context) (types.Info, error) {
+	env := types.Info{
+		Containers:         108,
+		ContainersRunning:  98,
+		ContainersStopped:  6,
+		ContainersPaused:   3,
+		OomKillDisable:     false,
+		SystemTime:         "2016-02-24T00:55:09.15073105-05:00",
+		NEventsListener:    0,
+		ID:                 "5WQQ:TFWR:FDNG:OKQ3:37Y4:FJWG:QIKK:623T:R3ME:QTKB:A7F7:OLHD",
+		Debug:              false,
+		LoggingDriver:      "json-file",
+		KernelVersion:      "4.3.0-1-amd64",
+		IndexServerAddress: "https://index.docker.io/v1/",
+		MemTotal:           3840757760,
+		Images:             199,
+		CPUCfsQuota:        true,
+		Name:               "absol",
+		SwapLimit:          false,
+		IPv4Forwarding:     true,
+		ExperimentalBuild:  false,
+		CPUCfsPeriod:       true,
+		RegistryConfig: &registry.ServiceConfig{
+			IndexConfigs: map[string]*registry.IndexInfo{
+				"docker.io": {
+					Name:     "docker.io",
+					Mirrors:  []string{},
+					Official: true,
+					Secure:   true,
+				},
+			}, InsecureRegistryCIDRs: []*registry.NetIPNet{{IP: []byte{127, 0, 0, 0}, Mask: []byte{255, 0, 0, 0}}}, Mirrors: []string{}},
+		OperatingSystem:   "Linux Mint LMDE (containerized)",
+		BridgeNfIptables:  true,
+		HTTPSProxy:        "",
+		Labels:            []string{},
+		MemoryLimit:       false,
+		DriverStatus:      [][2]string{{"Pool Name", "docker-8:1-1182287-pool"}, {"Pool Blocksize", "65.54 kB"}, {"Backing Filesystem", "extfs"}, {"Data file", "/dev/loop0"}, {"Metadata file", "/dev/loop1"}, {"Data Space Used", "17.3 GB"}, {"Data Space Total", "107.4 GB"}, {"Data Space Available", "36.53 GB"}, {"Metadata Space Used", "20.97 MB"}, {"Metadata Space Total", "2.147 GB"}, {"Metadata Space Available", "2.127 GB"}, {"Udev Sync Supported", "true"}, {"Deferred Removal Enabled", "false"}, {"Data loop file", "/var/lib/docker/devicemapper/devicemapper/data"}, {"Metadata loop file", "/var/lib/docker/devicemapper/devicemapper/metadata"}, {"Library Version", "1.02.115 (2016-01-25)"}},
+		NFd:               19,
+		HTTPProxy:         "",
+		Driver:            "devicemapper",
+		NGoroutines:       39,
+		NCPU:              4,
+		DockerRootDir:     "/var/lib/docker",
+		NoProxy:           "",
+		BridgeNfIP6tables: true,
+	}
+	return env, nil
+}
+
+func (d FakeDockerClient) ContainerList(octx context.Context, options types.ContainerListOptions) ([]types.Container, error) {
+	container1 := types.Container{
+		ID:      "e2173b9478a6ae55e237d4d74f8bbb753f0817192b5081334dc78476296b7dfb",
+		Names:   []string{"/etcd"},
+		Image:   "quay.io/coreos/etcd:v2.2.2",
+		Command: "/etcd -name etcd0 -advertise-client-urls http://localhost:2379 -listen-client-urls http://0.0.0.0:2379",
+		Created: 1455941930,
+		Status:  "Up 4 hours",
+		Ports: []types.Port{
+			types.Port{
+				PrivatePort: 7001,
+				PublicPort:  0,
+				Type:        "tcp",
+			},
+			types.Port{
+				PrivatePort: 4001,
+				PublicPort:  0,
+				Type:        "tcp",
+			},
+			types.Port{
+				PrivatePort: 2380,
+				PublicPort:  0,
+				Type:        "tcp",
+			},
+			types.Port{
+				PrivatePort: 2379,
+				PublicPort:  2379,
+				Type:        "tcp",
+				IP:          "0.0.0.0",
+			},
+		},
+		Labels: map[string]string{
+			"label1": "test_value_1",
+			"label2": "test_value_2",
+		},
+		SizeRw:     0,
+		SizeRootFs: 0,
+	}
+	container2 := types.Container{
+		ID:      "b7dfbb9478a6ae55e237d4d74f8bbb753f0817192b5081334dc78476296e2173",
+		Names:   []string{"/etcd2"},
+		Image:   "quay.io:4443/coreos/etcd:v2.2.2",
+		Command: "/etcd -name etcd2 -advertise-client-urls http://localhost:2379 -listen-client-urls http://0.0.0.0:2379",
+		Created: 1455941933,
+		Status:  "Up 4 hours",
+		Ports: []types.Port{
+			types.Port{
+				PrivatePort: 7002,
+				PublicPort:  0,
+				Type:        "tcp",
+			},
+			types.Port{
+				PrivatePort: 4002,
+				PublicPort:  0,
+				Type:        "tcp",
+			},
+			types.Port{
+				PrivatePort: 2381,
+				PublicPort:  0,
+				Type:        "tcp",
+			},
+			types.Port{
+				PrivatePort: 2382,
+				PublicPort:  2382,
+				Type:        "tcp",
+				IP:          "0.0.0.0",
+			},
+		},
+		Labels: map[string]string{
+			"label1": "test_value_1",
+			"label2": "test_value_2",
+		},
+		SizeRw:     0,
+		SizeRootFs: 0,
+	}
+
+	containers := []types.Container{container1, container2}
+	return containers, nil
+
+	//#{e6a96c84ca91a5258b7cb752579fb68826b68b49ff957487695cd4d13c343b44 titilambert/snmpsim /bin/sh -c 'snmpsimd --agent-udpv4-endpoint=0.0.0.0:31161 --process-user=root --process-group=user' 1455724831 Up 4 hours [{31161 31161 udp 0.0.0.0}] 0 0 [/snmp] map[]}]2016/02/24 01:05:01 Gathered metrics, (3s interval), from 1 inputs in 1.233836656s
+}
+
+func (d FakeDockerClient) ContainerStats(ctx context.Context, containerID string, stream bool) (types.ContainerStats, error) {
+	var stat types.ContainerStats
+	jsonStat := `{"read":"2016-02-24T11:42:27.472459608-05:00","memory_stats":{"stats":{},"limit":18935443456},"blkio_stats":{"io_service_bytes_recursive":[{"major":252,"minor":1,"op":"Read","value":753664},{"major":252,"minor":1,"op":"Write"},{"major":252,"minor":1,"op":"Sync"},{"major":252,"minor":1,"op":"Async","value":753664},{"major":252,"minor":1,"op":"Total","value":753664}],"io_serviced_recursive":[{"major":252,"minor":1,"op":"Read","value":26},{"major":252,"minor":1,"op":"Write"},{"major":252,"minor":1,"op":"Sync"},{"major":252,"minor":1,"op":"Async","value":26},{"major":252,"minor":1,"op":"Total","value":26}]},"cpu_stats":{"cpu_usage":{"percpu_usage":[17871,4959158,1646137,1231652,11829401,244656,369972,0],"usage_in_usermode":10000000,"total_usage":20298847},"system_cpu_usage":24052607520000000,"throttling_data":{}},"precpu_stats":{"cpu_usage":{"percpu_usage":[17871,4959158,1646137,1231652,11829401,244656,369972,0],"usage_in_usermode":10000000,"total_usage":20298847},"system_cpu_usage":24052599550000000,"throttling_data":{}}}`
+	stat.Body = ioutil.NopCloser(strings.NewReader(jsonStat))
+	return stat, nil
+}

plugins/inputs/dovecot/dovecot.go

@@ -12,7 +12,6 @@ import (
 	"time"
 
 	"github.com/influxdata/telegraf"
-	"github.com/influxdata/telegraf/internal/errchan"
 	"github.com/influxdata/telegraf/plugins/inputs"
 )
 
@@ -66,19 +65,18 @@ func (d *Dovecot) Gather(acc telegraf.Accumulator) error {
 	}
 
 	var wg sync.WaitGroup
-	errChan := errchan.New(len(d.Servers) * len(d.Filters))
 	for _, server := range d.Servers {
 		for _, filter := range d.Filters {
 			wg.Add(1)
 			go func(s string, f string) {
 				defer wg.Done()
-				errChan.C <- d.gatherServer(s, acc, d.Type, f)
+				acc.AddError(d.gatherServer(s, acc, d.Type, f))
 			}(server, filter)
 		}
 	}
 
 	wg.Wait()
-	return errChan.Error()
+	return nil
 }
 
 func (d *Dovecot) gatherServer(addr string, acc telegraf.Accumulator, qtype string, filter string) error {

plugins/inputs/elasticsearch/README.md

@@ -2,7 +2,8 @@
 
 The [elasticsearch](https://www.elastic.co/) plugin queries endpoints to obtain
 [node](https://www.elastic.co/guide/en/elasticsearch/reference/current/cluster-nodes-stats.html)
-and optionally [cluster](https://www.elastic.co/guide/en/elasticsearch/reference/current/cluster-health.html) stats.
+and optionally [cluster-health](https://www.elastic.co/guide/en/elasticsearch/reference/current/cluster-health.html)
+or [cluster-stats](https://www.elastic.co/guide/en/elasticsearch/reference/current/cluster-stats.html) metrics.
 
 ### Configuration:
 
@@ -14,13 +15,18 @@ and optionally [cluster](https://www.elastic.co/guide/en/elasticsearch/reference
   ## Timeout for HTTP requests to the elastic search server(s)
   http_timeout = "5s"
 
-  ## set local to false when you want to read the indices stats from all nodes
-  ## within the cluster
+  ## When local is true (the default), the node will read only its own stats.
+  ## Set local to false when you want to read the node stats from all nodes
+  ## of the cluster. 
   local = true
 
-  ## set cluster_health to true when you want to also obtain cluster level stats
+  ## Set cluster_health to true when you want to also obtain cluster health stats
   cluster_health = false
 
+  ## Set cluster_stats to true when you want to obtain cluster stats from the 
+  ## Master node. 
+  cluster_stats = false
+
   ## Optional SSL Config
   # ssl_ca = "/etc/telegraf/ca.pem"
   # ssl_cert = "/etc/telegraf/cert.pem"

plugins/inputs/elasticsearch/elasticsearch.go

@@ -4,21 +4,26 @@ import (
 	"encoding/json"
 	"fmt"
 	"net/http"
+	"regexp"
 	"sync"
 	"time"
 
 	"github.com/influxdata/telegraf"
 	"github.com/influxdata/telegraf/internal"
-	"github.com/influxdata/telegraf/internal/errchan"
 	"github.com/influxdata/telegraf/plugins/inputs"
 	jsonparser "github.com/influxdata/telegraf/plugins/parsers/json"
+	"io/ioutil"
+	"strings"
 )
 
+// mask for masking username/password from error messages
+var mask = regexp.MustCompile(`https?:\/\/\S+:\S+@`)
+
+// Nodestats are always generated, so simply define a constant for these endpoints
 const statsPath = "/_nodes/stats"
 const statsPathLocal = "/_nodes/_local/stats"
-const healthPath = "/_cluster/health"
 
-type node struct {
+type nodeStat struct {
 	Host       string            `json:"host"`
 	Name       string            `json:"name"`
 	Attributes map[string]string `json:"attributes"`
@@ -58,20 +63,41 @@ type indexHealth struct {
 	UnassignedShards    int    `json:"unassigned_shards"`
 }
 
+type clusterStats struct {
+	NodeName    string      `json:"node_name"`
+	ClusterName string      `json:"cluster_name"`
+	Status      string      `json:"status"`
+	Indices     interface{} `json:"indices"`
+	Nodes       interface{} `json:"nodes"`
+}
+
+type catMaster struct {
+	NodeID   string `json:"id"`
+	NodeIP   string `json:"ip"`
+	NodeName string `json:"node"`
+}
+
 const sampleConfig = `
   ## specify a list of one or more Elasticsearch servers
+  # you can add username and password to your url to use basic authentication:
+  # servers = ["http://user:pass@localhost:9200"]
   servers = ["http://localhost:9200"]
 
   ## Timeout for HTTP requests to the elastic search server(s)
   http_timeout = "5s"
 
-  ## set local to false when you want to read the indices stats from all nodes
-  ## within the cluster
+  ## When local is true (the default), the node will read only its own stats.
+  ## Set local to false when you want to read the node stats from all nodes
+  ## of the cluster.
   local = true
 
-  ## set cluster_health to true when you want to also obtain cluster level stats
+  ## Set cluster_health to true when you want to also obtain cluster health stats
   cluster_health = false
 
+  ## Set cluster_stats to true when you want to also obtain cluster stats from the
+  ## Master node.
+  cluster_stats = false
+
   ## Optional SSL Config
   # ssl_ca = "/etc/telegraf/ca.pem"
   # ssl_cert = "/etc/telegraf/cert.pem"
@@ -83,15 +109,18 @@ const sampleConfig = `
 // Elasticsearch is a plugin to read stats from one or many Elasticsearch
 // servers.
 type Elasticsearch struct {
-	Local              bool
-	Servers            []string
-	HttpTimeout        internal.Duration
-	ClusterHealth      bool
-	SSLCA              string `toml:"ssl_ca"`   // Path to CA file
-	SSLCert            string `toml:"ssl_cert"` // Path to host cert file
-	SSLKey             string `toml:"ssl_key"`  // Path to cert key file
-	InsecureSkipVerify bool   // Use SSL but skip chain & host verification
-	client             *http.Client
+	Local                   bool
+	Servers                 []string
+	HttpTimeout             internal.Duration
+	ClusterHealth           bool
+	ClusterStats            bool
+	SSLCA                   string `toml:"ssl_ca"`   // Path to CA file
+	SSLCert                 string `toml:"ssl_cert"` // Path to host cert file
+	SSLKey                  string `toml:"ssl_key"`  // Path to cert key file
+	InsecureSkipVerify      bool   // Use SSL but skip chain & host verification
+	client                  *http.Client
+	catMasterResponseTokens []string
+	isMaster                bool
 }
 
 // NewElasticsearch return a new instance of Elasticsearch
@@ -123,7 +152,6 @@ func (e *Elasticsearch) Gather(acc telegraf.Accumulator) error {
 		e.client = client
 	}
 
-	errChan := errchan.New(len(e.Servers))
 	var wg sync.WaitGroup
 	wg.Add(len(e.Servers))
 
@@ -136,18 +164,39 @@ func (e *Elasticsearch) Gather(acc telegraf.Accumulator) error {
 			} else {
 				url = s + statsPath
 			}
+			e.isMaster = false
+
+			if e.ClusterStats {
+				// get cat/master information here so NodeStats can determine
+				// whether this node is the Master
+				e.setCatMaster(s + "/_cat/master")
+			}
+
+			// Always gather node states
 			if err := e.gatherNodeStats(url, acc); err != nil {
-				errChan.C <- err
+				acc.AddError(fmt.Errorf(mask.ReplaceAllString(err.Error(), "http(s)://XXX:XXX@")))
 				return
 			}
+
 			if e.ClusterHealth {
-				e.gatherClusterStats(fmt.Sprintf("%s/_cluster/health?level=indices", s), acc)
+				url = s + "/_cluster/health?level=indices"
+				if err := e.gatherClusterHealth(url, acc); err != nil {
+					acc.AddError(fmt.Errorf(mask.ReplaceAllString(err.Error(), "http(s)://XXX:XXX@")))
+					return
+				}
+			}
+
+			if e.ClusterStats && e.isMaster {
+				if err := e.gatherClusterStats(s+"/_cluster/stats", acc); err != nil {
+					acc.AddError(fmt.Errorf(mask.ReplaceAllString(err.Error(), "http(s)://XXX:XXX@")))
+					return
+				}
 			}
 		}(serv, acc)
 	}
 
 	wg.Wait()
-	return errChan.Error()
+	return nil
 }
 
 func (e *Elasticsearch) createHttpClient() (*http.Client, error) {
@@ -169,12 +218,13 @@ func (e *Elasticsearch) createHttpClient() (*http.Client, error) {
 
 func (e *Elasticsearch) gatherNodeStats(url string, acc telegraf.Accumulator) error {
 	nodeStats := &struct {
-		ClusterName string           `json:"cluster_name"`
-		Nodes       map[string]*node `json:"nodes"`
+		ClusterName string               `json:"cluster_name"`
+		Nodes       map[string]*nodeStat `json:"nodes"`
 	}{}
-	if err := e.gatherData(url, nodeStats); err != nil {
+	if err := e.gatherJsonData(url, nodeStats); err != nil {
 		return err
 	}
+
 	for id, n := range nodeStats.Nodes {
 		tags := map[string]string{
 			"node_id":      id,
@@ -183,6 +233,11 @@ func (e *Elasticsearch) gatherNodeStats(url string, acc telegraf.Accumulator) er
 			"cluster_name": nodeStats.ClusterName,
 		}
 
+		if e.ClusterStats {
+			// check for master
+			e.isMaster = (id == e.catMasterResponseTokens[0])
+		}
+
 		for k, v := range n.Attributes {
 			tags["node_attribute_"+k] = v
 		}
@@ -202,6 +257,7 @@ func (e *Elasticsearch) gatherNodeStats(url string, acc telegraf.Accumulator) er
 		now := time.Now()
 		for p, s := range stats {
 			f := jsonparser.JSONFlattener{}
+			// parse Json, ignoring strings and bools
 			err := f.FlattenJSON("", s)
 			if err != nil {
 				return err
@@ -212,31 +268,31 @@ func (e *Elasticsearch) gatherNodeStats(url string, acc telegraf.Accumulator) er
 	return nil
 }
 
-func (e *Elasticsearch) gatherClusterStats(url string, acc telegraf.Accumulator) error {
-	clusterStats := &clusterHealth{}
-	if err := e.gatherData(url, clusterStats); err != nil {
+func (e *Elasticsearch) gatherClusterHealth(url string, acc telegraf.Accumulator) error {
+	healthStats := &clusterHealth{}
+	if err := e.gatherJsonData(url, healthStats); err != nil {
 		return err
 	}
 	measurementTime := time.Now()
 	clusterFields := map[string]interface{}{
-		"status":                clusterStats.Status,
-		"timed_out":             clusterStats.TimedOut,
-		"number_of_nodes":       clusterStats.NumberOfNodes,
-		"number_of_data_nodes":  clusterStats.NumberOfDataNodes,
-		"active_primary_shards": clusterStats.ActivePrimaryShards,
-		"active_shards":         clusterStats.ActiveShards,
-		"relocating_shards":     clusterStats.RelocatingShards,
-		"initializing_shards":   clusterStats.InitializingShards,
-		"unassigned_shards":     clusterStats.UnassignedShards,
+		"status":                healthStats.Status,
+		"timed_out":             healthStats.TimedOut,
+		"number_of_nodes":       healthStats.NumberOfNodes,
+		"number_of_data_nodes":  healthStats.NumberOfDataNodes,
+		"active_primary_shards": healthStats.ActivePrimaryShards,
+		"active_shards":         healthStats.ActiveShards,
+		"relocating_shards":     healthStats.RelocatingShards,
+		"initializing_shards":   healthStats.InitializingShards,
+		"unassigned_shards":     healthStats.UnassignedShards,
 	}
 	acc.AddFields(
 		"elasticsearch_cluster_health",
 		clusterFields,
-		map[string]string{"name": clusterStats.ClusterName},
+		map[string]string{"name": healthStats.ClusterName},
 		measurementTime,
 	)
 
-	for name, health := range clusterStats.Indices {
+	for name, health := range healthStats.Indices {
 		indexFields := map[string]interface{}{
 			"status":                health.Status,
 			"number_of_shards":      health.NumberOfShards,
@@ -257,7 +313,60 @@ func (e *Elasticsearch) gatherClusterStats(url string, acc telegraf.Accumulator)
 	return nil
 }
 
-func (e *Elasticsearch) gatherData(url string, v interface{}) error {
+func (e *Elasticsearch) gatherClusterStats(url string, acc telegraf.Accumulator) error {
+	clusterStats := &clusterStats{}
+	if err := e.gatherJsonData(url, clusterStats); err != nil {
+		return err
+	}
+	now := time.Now()
+	tags := map[string]string{
+		"node_name":    clusterStats.NodeName,
+		"cluster_name": clusterStats.ClusterName,
+		"status":       clusterStats.Status,
+	}
+
+	stats := map[string]interface{}{
+		"nodes":   clusterStats.Nodes,
+		"indices": clusterStats.Indices,
+	}
+
+	for p, s := range stats {
+		f := jsonparser.JSONFlattener{}
+		// parse json, including bools and strings
+		err := f.FullFlattenJSON("", s, true, true)
+		if err != nil {
+			return err
+		}
+		acc.AddFields("elasticsearch_clusterstats_"+p, f.Fields, tags, now)
+	}
+
+	return nil
+}
+
+func (e *Elasticsearch) setCatMaster(url string) error {
+	r, err := e.client.Get(url)
+	if err != nil {
+		return err
+	}
+	defer r.Body.Close()
+	if r.StatusCode != http.StatusOK {
+		// NOTE: we are not going to read/discard r.Body under the assumption we'd prefer
+		// to let the underlying transport close the connection and re-establish a new one for
+		// future calls.
+		return fmt.Errorf("status-code %d, expected %d", r.StatusCode, http.StatusOK)
+	}
+	response, err := ioutil.ReadAll(r.Body)
+
+	if err != nil {
+		return err
+	}
+
+	e.catMasterResponseTokens = strings.Split(string(response), " ")
+
+	return nil
+}
+
+func (e *Elasticsearch) gatherJsonData(url string, v interface{}) error {
 	r, err := e.client.Get(url)
 	if err != nil {
 		return err
@@ -270,9 +379,11 @@ func (e *Elasticsearch) gatherData(url string, v interface{}) error {
 		return fmt.Errorf("elasticsearch: API responded with status-code %d, expected %d",
 			r.StatusCode, http.StatusOK)
 	}
+
 	if err = json.NewDecoder(r.Body).Decode(v); err != nil {
 		return err
 	}
+
 	return nil
 }
 

plugins/inputs/elasticsearch/elasticsearch_test.go

@@ -8,6 +8,8 @@ import (
 
 	"github.com/influxdata/telegraf/testutil"
 
+	"fmt"
+	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
 
@@ -37,16 +39,13 @@ func (t *transportMock) RoundTrip(r *http.Request) (*http.Response, error) {
 func (t *transportMock) CancelRequest(_ *http.Request) {
 }
 
-func TestElasticsearch(t *testing.T) {
-	es := newElasticsearchWithClient()
-	es.Servers = []string{"http://example.com:9200"}
-	es.client.Transport = newTransportMock(http.StatusOK, statsResponse)
-
-	var acc testutil.Accumulator
-	if err := es.Gather(&acc); err != nil {
-		t.Fatal(err)
+func checkIsMaster(es *Elasticsearch, expected bool, t *testing.T) {
+	if es.isMaster != expected {
+		msg := fmt.Sprintf("IsMaster set incorrectly")
+		assert.Fail(t, msg)
 	}
-
+}
+func checkNodeStatsResult(t *testing.T, acc *testutil.Accumulator) {
 	tags := map[string]string{
 		"cluster_name":          "es-testcluster",
 		"node_attribute_master": "true",
@@ -55,25 +54,55 @@ func TestElasticsearch(t *testing.T) {
 		"node_host":             "test",
 	}
 
-	acc.AssertContainsTaggedFields(t, "elasticsearch_indices", indicesExpected, tags)
-	acc.AssertContainsTaggedFields(t, "elasticsearch_os", osExpected, tags)
-	acc.AssertContainsTaggedFields(t, "elasticsearch_process", processExpected, tags)
-	acc.AssertContainsTaggedFields(t, "elasticsearch_jvm", jvmExpected, tags)
-	acc.AssertContainsTaggedFields(t, "elasticsearch_thread_pool", threadPoolExpected, tags)
-	acc.AssertContainsTaggedFields(t, "elasticsearch_fs", fsExpected, tags)
-	acc.AssertContainsTaggedFields(t, "elasticsearch_transport", transportExpected, tags)
-	acc.AssertContainsTaggedFields(t, "elasticsearch_http", httpExpected, tags)
-	acc.AssertContainsTaggedFields(t, "elasticsearch_breakers", breakersExpected, tags)
+	acc.AssertContainsTaggedFields(t, "elasticsearch_indices", nodestatsIndicesExpected, tags)
+	acc.AssertContainsTaggedFields(t, "elasticsearch_os", nodestatsOsExpected, tags)
+	acc.AssertContainsTaggedFields(t, "elasticsearch_process", nodestatsProcessExpected, tags)
+	acc.AssertContainsTaggedFields(t, "elasticsearch_jvm", nodestatsJvmExpected, tags)
+	acc.AssertContainsTaggedFields(t, "elasticsearch_thread_pool", nodestatsThreadPoolExpected, tags)
+	acc.AssertContainsTaggedFields(t, "elasticsearch_fs", nodestatsFsExpected, tags)
+	acc.AssertContainsTaggedFields(t, "elasticsearch_transport", nodestatsTransportExpected, tags)
+	acc.AssertContainsTaggedFields(t, "elasticsearch_http", nodestatsHttpExpected, tags)
+	acc.AssertContainsTaggedFields(t, "elasticsearch_breakers", nodestatsBreakersExpected, tags)
 }
 
-func TestGatherClusterStats(t *testing.T) {
+func TestGather(t *testing.T) {
+	es := newElasticsearchWithClient()
+	es.Servers = []string{"http://example.com:9200"}
+	es.client.Transport = newTransportMock(http.StatusOK, nodeStatsResponse)
+
+	var acc testutil.Accumulator
+	if err := acc.GatherError(es.Gather); err != nil {
+		t.Fatal(err)
+	}
+
+	checkIsMaster(es, false, t)
+	checkNodeStatsResult(t, &acc)
+}
+
+func TestGatherNodeStats(t *testing.T) {
+	es := newElasticsearchWithClient()
+	es.Servers = []string{"http://example.com:9200"}
+	es.client.Transport = newTransportMock(http.StatusOK, nodeStatsResponse)
+
+	var acc testutil.Accumulator
+	if err := es.gatherNodeStats("junk", &acc); err != nil {
+		t.Fatal(err)
+	}
+
+	checkIsMaster(es, false, t)
+	checkNodeStatsResult(t, &acc)
+}
+
+func TestGatherClusterHealth(t *testing.T) {
 	es := newElasticsearchWithClient()
 	es.Servers = []string{"http://example.com:9200"}
 	es.ClusterHealth = true
-	es.client.Transport = newTransportMock(http.StatusOK, clusterResponse)
+	es.client.Transport = newTransportMock(http.StatusOK, clusterHealthResponse)
 
 	var acc testutil.Accumulator
-	require.NoError(t, es.Gather(&acc))
+	require.NoError(t, es.gatherClusterHealth("junk", &acc))
+
+	checkIsMaster(es, false, t)
 
 	acc.AssertContainsTaggedFields(t, "elasticsearch_cluster_health",
 		clusterHealthExpected,
@@ -88,6 +117,77 @@ func TestGatherClusterStats(t *testing.T) {
 		map[string]string{"index": "v2"})
 }
 
+func TestGatherClusterStatsMaster(t *testing.T) {
+	// This needs multiple steps to replicate the multiple calls internally.
+	es := newElasticsearchWithClient()
+	es.ClusterStats = true
+	es.Servers = []string{"http://example.com:9200"}
+
+	// first get catMaster
+	es.client.Transport = newTransportMock(http.StatusOK, IsMasterResult)
+	require.NoError(t, es.setCatMaster("junk"))
+
+	IsMasterResultTokens := strings.Split(string(IsMasterResult), " ")
+	if es.catMasterResponseTokens[0] != IsMasterResultTokens[0] {
+		msg := fmt.Sprintf("catmaster is incorrect")
+		assert.Fail(t, msg)
+	}
+
+	// now get node status, which determines whether we're master
+	var acc testutil.Accumulator
+	es.Local = true
+	es.client.Transport = newTransportMock(http.StatusOK, nodeStatsResponse)
+	if err := es.gatherNodeStats("junk", &acc); err != nil {
+		t.Fatal(err)
+	}
+
+	checkIsMaster(es, true, t)
+	checkNodeStatsResult(t, &acc)
+
+	// now test the clusterstats method
+	es.client.Transport = newTransportMock(http.StatusOK, clusterStatsResponse)
+	require.NoError(t, es.gatherClusterStats("junk", &acc))
+
+	tags := map[string]string{
+		"cluster_name": "es-testcluster",
+		"node_name":    "test.host.com",
+		"status":       "red",
+	}
+
+	acc.AssertContainsTaggedFields(t, "elasticsearch_clusterstats_nodes", clusterstatsNodesExpected, tags)
+	acc.AssertContainsTaggedFields(t, "elasticsearch_clusterstats_indices", clusterstatsIndicesExpected, tags)
+}
+
+func TestGatherClusterStatsNonMaster(t *testing.T) {
+	// This needs multiple steps to replicate the multiple calls internally.
+	es := newElasticsearchWithClient()
+	es.ClusterStats = true
+	es.Servers = []string{"http://example.com:9200"}
+
+	// first get catMaster
+	es.client.Transport = newTransportMock(http.StatusOK, IsNotMasterResult)
+	require.NoError(t, es.setCatMaster("junk"))
+
+	IsNotMasterResultTokens := strings.Split(string(IsNotMasterResult), " ")
+	if es.catMasterResponseTokens[0] != IsNotMasterResultTokens[0] {
+		msg := fmt.Sprintf("catmaster is incorrect")
+		assert.Fail(t, msg)
+	}
+
+	// now get node status, which determines whether we're master
+	var acc testutil.Accumulator
+	es.Local = true
+	es.client.Transport = newTransportMock(http.StatusOK, nodeStatsResponse)
+	if err := es.gatherNodeStats("junk", &acc); err != nil {
+		t.Fatal(err)
+	}
+
+	// ensure flag is clear so Cluster Stats would not be done
+	checkIsMaster(es, false, t)
+	checkNodeStatsResult(t, &acc)
+
+}
+
 func newElasticsearchWithClient() *Elasticsearch {
 	es := NewElasticsearch()
 	es.client = &http.Client{}

plugins/inputs/elasticsearch/testdata_test.go

@@ -1,6 +1,6 @@
 package elasticsearch
 
-const clusterResponse = `
+const clusterHealthResponse = `
 {
    "cluster_name": "elasticsearch_telegraf",
    "status": "green",
@@ -71,7 +71,7 @@ var v2IndexExpected = map[string]interface{}{
 	"unassigned_shards":     20,
 }
 
-const statsResponse = `
+const nodeStatsResponse = `
 {
   "cluster_name": "es-testcluster",
   "nodes": {
@@ -489,7 +489,7 @@ const statsResponse = `
 }
 `
 
-var indicesExpected = map[string]interface{}{
+var nodestatsIndicesExpected = map[string]interface{}{
 	"id_cache_memory_size_in_bytes":             float64(0),
 	"completion_size_in_bytes":                  float64(0),
 	"suggest_total":                             float64(0),
@@ -561,7 +561,7 @@ var indicesExpected = map[string]interface{}{
 	"segments_fixed_bit_set_memory_in_bytes":    float64(0),
 }
 
-var osExpected = map[string]interface{}{
+var nodestatsOsExpected = map[string]interface{}{
 	"load_average_0":           float64(0.01),
 	"load_average_1":           float64(0.04),
 	"load_average_2":           float64(0.05),
@@ -576,7 +576,7 @@ var osExpected = map[string]interface{}{
 	"mem_used_in_bytes":        float64(1621868544),
 }
 
-var processExpected = map[string]interface{}{
+var nodestatsProcessExpected = map[string]interface{}{
 	"mem_total_virtual_in_bytes": float64(4747890688),
 	"timestamp":                  float64(1436460392945),
 	"open_file_descriptors":      float64(160),
@@ -586,7 +586,7 @@ var processExpected = map[string]interface{}{
 	"cpu_user_in_millis":         float64(13610),
 }
 
-var jvmExpected = map[string]interface{}{
+var nodestatsJvmExpected = map[string]interface{}{
 	"timestamp":                                     float64(1436460392945),
 	"uptime_in_millis":                              float64(202245),
 	"mem_non_heap_used_in_bytes":                    float64(39634576),
@@ -621,7 +621,7 @@ var jvmExpected = map[string]interface{}{
 	"buffer_pools_mapped_total_capacity_in_bytes":   float64(0),
 }
 
-var threadPoolExpected = map[string]interface{}{
+var nodestatsThreadPoolExpected = map[string]interface{}{
 	"merge_threads":                 float64(6),
 	"merge_queue":                   float64(4),
 	"merge_active":                  float64(5),
@@ -726,7 +726,7 @@ var threadPoolExpected = map[string]interface{}{
 	"flush_completed":               float64(3),
 }
 
-var fsExpected = map[string]interface{}{
+var nodestatsFsExpected = map[string]interface{}{
 	"data_0_total_in_bytes":     float64(19507089408),
 	"data_0_free_in_bytes":      float64(16909316096),
 	"data_0_available_in_bytes": float64(15894814720),
@@ -736,7 +736,7 @@ var fsExpected = map[string]interface{}{
 	"total_total_in_bytes":      float64(19507089408),
 }
 
-var transportExpected = map[string]interface{}{
+var nodestatsTransportExpected = map[string]interface{}{
 	"server_open":      float64(13),
 	"rx_count":         float64(6),
 	"rx_size_in_bytes": float64(1380),
@@ -744,12 +744,12 @@ var transportExpected = map[string]interface{}{
 	"tx_size_in_bytes": float64(1380),
 }
 
-var httpExpected = map[string]interface{}{
+var nodestatsHttpExpected = map[string]interface{}{
 	"current_open": float64(3),
 	"total_opened": float64(3),
 }
 
-var breakersExpected = map[string]interface{}{
+var nodestatsBreakersExpected = map[string]interface{}{
 	"fielddata_estimated_size_in_bytes": float64(0),
 	"fielddata_overhead":                float64(1.03),
 	"fielddata_tripped":                 float64(0),
@@ -763,3 +763,273 @@ var breakersExpected = map[string]interface{}{
 	"parent_limit_size_in_bytes":        float64(727213670),
 	"parent_estimated_size_in_bytes":    float64(0),
 }
+
+const clusterStatsResponse = `
+{
+   "host":"ip-10-0-1-214",
+   "log_type":"metrics",
+   "timestamp":1475767451229,
+   "log_level":"INFO",
+   "node_name":"test.host.com",
+   "cluster_name":"es-testcluster",
+   "status":"red",
+   "indices":{
+      "count":1,
+      "shards":{
+         "total":4,
+         "primaries":4,
+         "replication":0.0,
+         "index":{
+            "shards":{
+               "min":4,
+               "max":4,
+               "avg":4.0
+            },
+            "primaries":{
+               "min":4,
+               "max":4,
+               "avg":4.0
+            },
+            "replication":{
+               "min":0.0,
+               "max":0.0,
+               "avg":0.0
+            }
+         }
+      },
+      "docs":{
+         "count":4,
+         "deleted":0
+      },
+      "store":{
+         "size_in_bytes":17084,
+         "throttle_time_in_millis":0
+      },
+      "fielddata":{
+         "memory_size_in_bytes":0,
+         "evictions":0
+      },
+      "query_cache":{
+         "memory_size_in_bytes":0,
+         "total_count":0,
+         "hit_count":0,
+         "miss_count":0,
+         "cache_size":0,
+         "cache_count":0,
+         "evictions":0
+      },
+      "completion":{
+         "size_in_bytes":0
+      },
+      "segments":{
+         "count":4,
+         "memory_in_bytes":11828,
+         "terms_memory_in_bytes":8932,
+         "stored_fields_memory_in_bytes":1248,
+         "term_vectors_memory_in_bytes":0,
+         "norms_memory_in_bytes":1280,
+         "doc_values_memory_in_bytes":368,
+         "index_writer_memory_in_bytes":0,
+         "index_writer_max_memory_in_bytes":2048000,
+         "version_map_memory_in_bytes":0,
+         "fixed_bit_set_memory_in_bytes":0
+      },
+      "percolate":{
+         "total":0,
+         "time_in_millis":0,
+         "current":0,
+         "memory_size_in_bytes":-1,
+         "memory_size":"-1b",
+         "queries":0
+      }
+   },
+   "nodes":{
+      "count":{
+         "total":1,
+         "master_only":0,
+         "data_only":0,
+         "master_data":1,
+         "client":0
+      },
+      "versions":[
+         {
+         "version": "2.3.3"
+         }
+      ],
+      "os":{
+         "available_processors":1,
+         "allocated_processors":1,
+         "mem":{
+            "total_in_bytes":593301504
+         },
+         "names":[
+            {
+               "name":"Linux",
+               "count":1
+            }
+         ]
+      },
+      "process":{
+         "cpu":{
+            "percent":0
+         },
+         "open_file_descriptors":{
+            "min":145,
+            "max":145,
+            "avg":145
+         }
+      },
+      "jvm":{
+         "max_uptime_in_millis":11580527,
+         "versions":[
+            {
+               "version":"1.8.0_101",
+               "vm_name":"OpenJDK 64-Bit Server VM",
+               "vm_version":"25.101-b13",
+               "vm_vendor":"Oracle Corporation",
+               "count":1
+            }
+         ],
+         "mem":{
+            "heap_used_in_bytes":70550288,
+            "heap_max_in_bytes":1065025536
+         },
+         "threads":30
+      },
+      "fs":{
+         "total_in_bytes":8318783488,
+         "free_in_bytes":6447439872,
+         "available_in_bytes":6344785920
+      },
+      "plugins":[
+         {
+            "name":"cloud-aws",
+            "version":"2.3.3",
+            "description":"The Amazon Web Service (AWS) Cloud plugin allows to use AWS API for the unicast discovery mechanism and add S3 repositories.",
+            "jvm":true,
+            "classname":"org.elasticsearch.plugin.cloud.aws.CloudAwsPlugin",
+            "isolated":true,
+            "site":false
+         },
+         {
+            "name":"kopf",
+            "version":"2.0.1",
+            "description":"kopf - simple web administration tool for Elasticsearch",
+            "url":"/_plugin/kopf/",
+            "jvm":false,
+            "site":true
+         },
+         {
+            "name":"tr-metrics",
+            "version":"7bd5b4b",
+            "description":"Logs cluster and node stats for performance monitoring.",
+            "jvm":true,
+            "classname":"com.trgr.elasticsearch.plugin.metrics.MetricsPlugin",
+            "isolated":true,
+            "site":false
+         }
+      ]
+   }
+}
+`
+
+var clusterstatsIndicesExpected = map[string]interface{}{
+	"completion_size_in_bytes":                  float64(0),
+	"count":                                     float64(1),
+	"docs_count":                                float64(4),
+	"docs_deleted":                              float64(0),
+	"fielddata_evictions":                       float64(0),
+	"fielddata_memory_size_in_bytes":            float64(0),
+	"percolate_current":                         float64(0),
+	"percolate_memory_size_in_bytes":            float64(-1),
+	"percolate_queries":                         float64(0),
+	"percolate_time_in_millis":                  float64(0),
+	"percolate_total":                           float64(0),
+	"percolate_memory_size":                     "-1b",
+	"query_cache_cache_count":                   float64(0),
+	"query_cache_cache_size":                    float64(0),
+	"query_cache_evictions":                     float64(0),
+	"query_cache_hit_count":                     float64(0),
+	"query_cache_memory_size_in_bytes":          float64(0),
+	"query_cache_miss_count":                    float64(0),
+	"query_cache_total_count":                   float64(0),
+	"segments_count":                            float64(4),
+	"segments_doc_values_memory_in_bytes":       float64(368),
+	"segments_fixed_bit_set_memory_in_bytes":    float64(0),
+	"segments_index_writer_max_memory_in_bytes": float64(2.048e+06),
+	"segments_index_writer_memory_in_bytes":     float64(0),
+	"segments_memory_in_bytes":                  float64(11828),
+	"segments_norms_memory_in_bytes":            float64(1280),
+	"segments_stored_fields_memory_in_bytes":    float64(1248),
+	"segments_term_vectors_memory_in_bytes":     float64(0),
+	"segments_terms_memory_in_bytes":            float64(8932),
+	"segments_version_map_memory_in_bytes":      float64(0),
+	"shards_index_primaries_avg":                float64(4),
+	"shards_index_primaries_max":                float64(4),
+	"shards_index_primaries_min":                float64(4),
+	"shards_index_replication_avg":              float64(0),
+	"shards_index_replication_max":              float64(0),
+	"shards_index_replication_min":              float64(0),
+	"shards_index_shards_avg":                   float64(4),
+	"shards_index_shards_max":                   float64(4),
+	"shards_index_shards_min":                   float64(4),
+	"shards_primaries":                          float64(4),
+	"shards_replication":                        float64(0),
+	"shards_total":                              float64(4),
+	"store_size_in_bytes":                       float64(17084),
+	"store_throttle_time_in_millis":             float64(0),
+}
+
+var clusterstatsNodesExpected = map[string]interface{}{
+	"count_client":                      float64(0),
+	"count_data_only":                   float64(0),
+	"count_master_data":                 float64(1),
+	"count_master_only":                 float64(0),
+	"count_total":                       float64(1),
+	"fs_available_in_bytes":             float64(6.34478592e+09),
+	"fs_free_in_bytes":                  float64(6.447439872e+09),
+	"fs_total_in_bytes":                 float64(8.318783488e+09),
+	"jvm_max_uptime_in_millis":          float64(1.1580527e+07),
+	"jvm_mem_heap_max_in_bytes":         float64(1.065025536e+09),
+	"jvm_mem_heap_used_in_bytes":        float64(7.0550288e+07),
+	"jvm_threads":                       float64(30),
+	"jvm_versions_0_count":              float64(1),
+	"jvm_versions_0_version":            "1.8.0_101",
+	"jvm_versions_0_vm_name":            "OpenJDK 64-Bit Server VM",
+	"jvm_versions_0_vm_vendor":          "Oracle Corporation",
+	"jvm_versions_0_vm_version":         "25.101-b13",
+	"os_allocated_processors":           float64(1),
+	"os_available_processors":           float64(1),
+	"os_mem_total_in_bytes":             float64(5.93301504e+08),
+	"os_names_0_count":                  float64(1),
+	"os_names_0_name":                   "Linux",
+	"process_cpu_percent":               float64(0),
+	"process_open_file_descriptors_avg": float64(145),
+	"process_open_file_descriptors_max": float64(145),
+	"process_open_file_descriptors_min": float64(145),
+	"versions_0_version":                "2.3.3",
+	"plugins_0_classname":               "org.elasticsearch.plugin.cloud.aws.CloudAwsPlugin",
+	"plugins_0_description":             "The Amazon Web Service (AWS) Cloud plugin allows to use AWS API for the unicast discovery mechanism and add S3 repositories.",
+	"plugins_0_isolated":                true,
+	"plugins_0_jvm":                     true,
+	"plugins_0_name":                    "cloud-aws",
+	"plugins_0_site":                    false,
+	"plugins_0_version":                 "2.3.3",
+	"plugins_1_description":             "kopf - simple web administration tool for Elasticsearch",
+	"plugins_1_jvm":                     false,
+	"plugins_1_name":                    "kopf",
+	"plugins_1_site":                    true,
+	"plugins_1_url":                     "/_plugin/kopf/",
+	"plugins_1_version":                 "2.0.1",
+	"plugins_2_classname":               "com.trgr.elasticsearch.plugin.metrics.MetricsPlugin",
+	"plugins_2_description":             "Logs cluster and node stats for performance monitoring.",
+	"plugins_2_isolated":                true,
+	"plugins_2_jvm":                     true,
+	"plugins_2_name":                    "tr-metrics",
+	"plugins_2_site":                    false,
+	"plugins_2_version":                 "7bd5b4b",
+}
+
+const IsMasterResult = "SDFsfSDFsdfFSDSDfSFDSDF 10.206.124.66 10.206.124.66 test.host.com "
+
+const IsNotMasterResult = "junk 10.206.124.66 10.206.124.66 test.junk.com "

plugins/inputs/exec/exec.go

@@ -11,11 +11,10 @@ import (
 	"syscall"
 	"time"
 
-	"github.com/gonuts/go-shellquote"
+	"github.com/kballard/go-shellquote"
 
 	"github.com/influxdata/telegraf"
 	"github.com/influxdata/telegraf/internal"
-	"github.com/influxdata/telegraf/internal/errchan"
 	"github.com/influxdata/telegraf/plugins/inputs"
 	"github.com/influxdata/telegraf/plugins/parsers"
 	"github.com/influxdata/telegraf/plugins/parsers/nagios"
@@ -36,7 +35,7 @@ const sampleConfig = `
   name_suffix = "_mycollector"
 
   ## Data format to consume.
-  ## Each data format has it's own unique set of configuration options, read
+  ## Each data format has its own unique set of configuration options, read
   ## more about them here:
   ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
   data_format = "influx"
@@ -49,8 +48,7 @@ type Exec struct {
 
 	parser parsers.Parser
 
-	runner  Runner
-	errChan chan error
+	runner Runner
 }
 
 func NewExec() *Exec {
@@ -150,13 +148,13 @@ func (e *Exec) ProcessCommand(command string, acc telegraf.Accumulator, wg *sync
 
 	out, err := e.runner.Run(e, command, acc)
 	if err != nil {
-		e.errChan <- err
+		acc.AddError(err)
 		return
 	}
 
 	metrics, err := e.parser.Parse(out)
 	if err != nil {
-		e.errChan <- err
+		acc.AddError(err)
 	} else {
 		for _, metric := range metrics {
 			acc.AddFields(metric.Name(), metric.Fields(), metric.Tags(), metric.Time())
@@ -193,7 +191,8 @@ func (e *Exec) Gather(acc telegraf.Accumulator) error {
 
 		matches, err := filepath.Glob(cmdAndArgs[0])
 		if err != nil {
-			return err
+			acc.AddError(err)
+			continue
 		}
 
 		if len(matches) == 0 {
@@ -214,15 +213,12 @@ func (e *Exec) Gather(acc telegraf.Accumulator) error {
 		}
 	}
 
-	errChan := errchan.New(len(commands))
-	e.errChan = errChan.C
-
 	wg.Add(len(commands))
 	for _, command := range commands {
 		go e.ProcessCommand(command, acc, &wg)
 	}
 	wg.Wait()
-	return errChan.Error()
+	return nil
 }
 
 func init() {

plugins/inputs/exec/exec_test.go

@@ -36,7 +36,9 @@ const malformedJson = `
     "status": "green",
 `
 
-const lineProtocol = "cpu,host=foo,datacenter=us-east usage_idle=99,usage_busy=1"
+const lineProtocol = "cpu,host=foo,datacenter=us-east usage_idle=99,usage_busy=1\n"
+const lineProtocolEmpty = ""
+const lineProtocolShort = "ab"
 
 const lineProtocolMulti = `
 cpu,cpu=cpu0,host=foo,datacenter=us-east usage_idle=99,usage_busy=1
@@ -99,7 +101,7 @@ func TestExec(t *testing.T) {
 	}
 
 	var acc testutil.Accumulator
-	err := e.Gather(&acc)
+	err := acc.GatherError(e.Gather)
 	require.NoError(t, err)
 	assert.Equal(t, acc.NFields(), 8, "non-numeric measurements should be ignored")
 
@@ -125,8 +127,7 @@ func TestExecMalformed(t *testing.T) {
 	}
 
 	var acc testutil.Accumulator
-	err := e.Gather(&acc)
-	require.Error(t, err)
+	require.Error(t, acc.GatherError(e.Gather))
 	assert.Equal(t, acc.NFields(), 0, "No new points should have been added")
 }
 
@@ -139,8 +140,7 @@ func TestCommandError(t *testing.T) {
 	}
 
 	var acc testutil.Accumulator
-	err := e.Gather(&acc)
-	require.Error(t, err)
+	require.Error(t, acc.GatherError(e.Gather))
 	assert.Equal(t, acc.NFields(), 0, "No new points should have been added")
 }
 
@@ -153,8 +153,7 @@ func TestLineProtocolParse(t *testing.T) {
 	}
 
 	var acc testutil.Accumulator
-	err := e.Gather(&acc)
-	require.NoError(t, err)
+	require.NoError(t, acc.GatherError(e.Gather))
 
 	fields := map[string]interface{}{
 		"usage_idle": float64(99),
@@ -167,6 +166,33 @@ func TestLineProtocolParse(t *testing.T) {
 	acc.AssertContainsTaggedFields(t, "cpu", fields, tags)
 }
 
+func TestLineProtocolEmptyParse(t *testing.T) {
+	parser, _ := parsers.NewInfluxParser()
+	e := &Exec{
+		runner:   newRunnerMock([]byte(lineProtocolEmpty), nil),
+		Commands: []string{"line-protocol"},
+		parser:   parser,
+	}
+
+	var acc testutil.Accumulator
+	err := e.Gather(&acc)
+	require.NoError(t, err)
+}
+
+func TestLineProtocolShortParse(t *testing.T) {
+	parser, _ := parsers.NewInfluxParser()
+	e := &Exec{
+		runner:   newRunnerMock([]byte(lineProtocolShort), nil),
+		Commands: []string{"line-protocol"},
+		parser:   parser,
+	}
+
+	var acc testutil.Accumulator
+	err := acc.GatherError(e.Gather)
+	require.Error(t, err)
+	assert.Contains(t, err.Error(), "buffer too short", "A buffer too short error was expected")
+}
+
 func TestLineProtocolParseMultiple(t *testing.T) {
 	parser, _ := parsers.NewInfluxParser()
 	e := &Exec{
@@ -176,7 +202,7 @@ func TestLineProtocolParseMultiple(t *testing.T) {
 	}
 
 	var acc testutil.Accumulator
-	err := e.Gather(&acc)
+	err := acc.GatherError(e.Gather)
 	require.NoError(t, err)
 
 	fields := map[string]interface{}{
@@ -202,7 +228,7 @@ func TestExecCommandWithGlob(t *testing.T) {
 	e.SetParser(parser)
 
 	var acc testutil.Accumulator
-	err := e.Gather(&acc)
+	err := acc.GatherError(e.Gather)
 	require.NoError(t, err)
 
 	fields := map[string]interface{}{
@@ -218,7 +244,7 @@ func TestExecCommandWithoutGlob(t *testing.T) {
 	e.SetParser(parser)
 
 	var acc testutil.Accumulator
-	err := e.Gather(&acc)
+	err := acc.GatherError(e.Gather)
 	require.NoError(t, err)
 
 	fields := map[string]interface{}{
@@ -234,7 +260,7 @@ func TestExecCommandWithoutGlobAndPath(t *testing.T) {
 	e.SetParser(parser)
 
 	var acc testutil.Accumulator
-	err := e.Gather(&acc)
+	err := acc.GatherError(e.Gather)
 	require.NoError(t, err)
 
 	fields := map[string]interface{}{

plugins/inputs/filestat/filestat.go

@@ -4,6 +4,7 @@ import (
 	"crypto/md5"
 	"fmt"
 	"io"
+	"log"
 	"os"
 
 	"github.com/influxdata/telegraf"
@@ -47,7 +48,6 @@ func (_ *FileStat) Description() string {
 func (_ *FileStat) SampleConfig() string { return sampleConfig }
 
 func (f *FileStat) Gather(acc telegraf.Accumulator) error {
-	var errS string
 	var err error
 
 	for _, filepath := range f.Files {
@@ -55,7 +55,7 @@ func (f *FileStat) Gather(acc telegraf.Accumulator) error {
 		g, ok := f.globs[filepath]
 		if !ok {
 			if g, err = globpath.Compile(filepath); err != nil {
-				errS += err.Error() + " "
+				acc.AddError(err)
 				continue
 			}
 			f.globs[filepath] = g
@@ -78,14 +78,20 @@ func (f *FileStat) Gather(acc telegraf.Accumulator) error {
 				"file": fileName,
 			}
 			fields := map[string]interface{}{
-				"exists":     int64(1),
-				"size_bytes": fileInfo.Size(),
+				"exists": int64(1),
+			}
+
+			if fileInfo == nil {
+				log.Printf("E! Unable to get info for file [%s], possible permissions issue",
+					fileName)
+			} else {
+				fields["size_bytes"] = fileInfo.Size()
 			}
 
 			if f.Md5 {
 				md5, err := getMd5(fileName)
 				if err != nil {
-					errS += err.Error() + " "
+					acc.AddError(err)
 				} else {
 					fields["md5_sum"] = md5
 				}
@@ -95,9 +101,6 @@ func (f *FileStat) Gather(acc telegraf.Accumulator) error {
 		}
 	}
 
-	if errS != "" {
-		return fmt.Errorf(errS)
-	}
 	return nil
 }
 

plugins/inputs/filestat/filestat_test.go

@@ -19,7 +19,7 @@ func TestGatherNoMd5(t *testing.T) {
 	}
 
 	acc := testutil.Accumulator{}
-	fs.Gather(&acc)
+	acc.GatherError(fs.Gather)
 
 	tags1 := map[string]string{
 		"file": dir + "log1.log",
@@ -59,7 +59,7 @@ func TestGatherExplicitFiles(t *testing.T) {
 	}
 
 	acc := testutil.Accumulator{}
-	fs.Gather(&acc)
+	acc.GatherError(fs.Gather)
 
 	tags1 := map[string]string{
 		"file": dir + "log1.log",
@@ -99,7 +99,7 @@ func TestGatherGlob(t *testing.T) {
 	}
 
 	acc := testutil.Accumulator{}
-	fs.Gather(&acc)
+	acc.GatherError(fs.Gather)
 
 	tags1 := map[string]string{
 		"file": dir + "log1.log",
@@ -131,7 +131,7 @@ func TestGatherSuperAsterisk(t *testing.T) {
 	}
 
 	acc := testutil.Accumulator{}
-	fs.Gather(&acc)
+	acc.GatherError(fs.Gather)
 
 	tags1 := map[string]string{
 		"file": dir + "log1.log",

plugins/inputs/graylog/graylog.go

@@ -4,7 +4,6 @@ import (
 	"bytes"
 	"encoding/base64"
 	"encoding/json"
-	"errors"
 	"fmt"
 	"io/ioutil"
 	"net"
@@ -149,31 +148,17 @@ func (h *GrayLog) Gather(acc telegraf.Accumulator) error {
 		h.client.SetHTTPClient(client)
 	}
 
-	errorChannel := make(chan error, len(h.Servers))
-
 	for _, server := range h.Servers {
 		wg.Add(1)
 		go func(server string) {
 			defer wg.Done()
-			if err := h.gatherServer(acc, server); err != nil {
-				errorChannel <- err
-			}
+			acc.AddError(h.gatherServer(acc, server))
 		}(server)
 	}
 
 	wg.Wait()
-	close(errorChannel)
 
-	// Get all errors and return them as one giant error
-	errorStrings := []string{}
-	for err := range errorChannel {
-		errorStrings = append(errorStrings, err.Error())
-	}
-
-	if len(errorStrings) == 0 {
-		return nil
-	}
-	return errors.New(strings.Join(errorStrings, "\n"))
+	return nil
 }
 
 // Gathers data from a particular server

plugins/inputs/graylog/graylog_test.go

@@ -157,7 +157,7 @@ func TestNormalResponse(t *testing.T) {
 
 	for _, service := range graylog {
 		var acc testutil.Accumulator
-		err := service.Gather(&acc)
+		err := acc.GatherError(service.Gather)
 		require.NoError(t, err)
 		for k, v := range expectedFields {
 			acc.AssertContainsTaggedFields(t, k, v, validTags[k])
@@ -170,9 +170,9 @@ func TestHttpJson500(t *testing.T) {
 	graylog := genMockGrayLog(validJSON, 500)
 
 	var acc testutil.Accumulator
-	err := graylog[0].Gather(&acc)
+	err := acc.GatherError(graylog[0].Gather)
 
-	assert.NotNil(t, err)
+	assert.Error(t, err)
 	assert.Equal(t, 0, acc.NFields())
 }
 
@@ -181,9 +181,9 @@ func TestHttpJsonBadJson(t *testing.T) {
 	graylog := genMockGrayLog(invalidJSON, 200)
 
 	var acc testutil.Accumulator
-	err := graylog[0].Gather(&acc)
+	err := acc.GatherError(graylog[0].Gather)
 
-	assert.NotNil(t, err)
+	assert.Error(t, err)
 	assert.Equal(t, 0, acc.NFields())
 }
 
@@ -192,8 +192,8 @@ func TestHttpJsonEmptyResponse(t *testing.T) {
 	graylog := genMockGrayLog(empty, 200)
 
 	var acc testutil.Accumulator
-	err := graylog[0].Gather(&acc)
+	err := acc.GatherError(graylog[0].Gather)
 
-	assert.NotNil(t, err)
+	assert.Error(t, err)
 	assert.Equal(t, 0, acc.NFields())
 }