By the time the aggregator.run() was called about 600ms already passed since setting now which was skewing up the aggregation intervals and skipping metrics.
Adjust logic in functions responsible for passing metrics in order to be able
to process them correctly in case where pass and drop are defined together.
Looping the metrics back through the same channel could result in a
deadlock, by using a new channel and locking the processor we can ensure
that all stages can make continual progress.
It is not possible to encode a measurement, tag, or field whose last
character is a backslash due to it being an unescapable character.
Because the tight coupling between line protocol and the internal metric
model, prevent metrics like this from being created.
Measurements with a trailing slash are not allowed and the point will be
dropped. Tags and fields with a trailing a slash will be dropped from
the point.
main reasons behind this:
- make adding/removing tags cheap
- make adding/removing fields cheap
- make parsing cheaper
- make parse -> decorate -> write out bytes metric flow much faster
Refactor serializer to use byte buffer
in this commit:
- chunks out the http request body to avoid making very large
allocations.
- establishes a limit for the maximum http request body size that the
listener will accept.
- utilizes a pool of byte buffers to reduce GC pressure.
in this commit:
- centralize logging output handler.
- set global Info/Debug/Error log levels based on config file or flags.
- remove per-plugin debug arg handling.
- add a I!, D!, or E! to every log message.
- add configuration option to specify where to send logs.
closes#1786
started working on this with the idea of fixing #1623, although I
realized that this was actually just a documentation issue around
a toml eccentricity.
closes#1623
this will now be the default config file location on windows, basically
it is the windows equivalent of /etc/telegraf/telegraf.conf
also updating the changelog
closes#1543
* Allow for TLS connections to ElasticSearch
Extremely similar implementation to the HTTP JSON module's
implementation of the same code.
* Changelog update
Adding precision rounding to the accumulator. This means that now every
input metric will get rounded at collection, rather than at write (and
only for the influxdb output).
This feature is disabled for service inputs, because service inputs
should be in control of their own timestamps & precisions.
use a common function between collection_jitter and flush_jitter. which
creates the same behavior between the two options.
going forward, both jitters will be random sleeps that get re-evaluated
at runtime for every interval (previously only collection_jitter did
this)
also fixes behavior so that both jitters will exit in the event of a
process exit.
closes#1296
* Use shared AWS credential configuration.
* Cloudwatch dimension wilcards
* Allow configuring cache_ttl for cloudwatch metrics.
* Allow for wildcard in dimension values to select all available metrics.
* Use internal.Duration for CacheTTL and go fmt
* Refactor to not use embedded structs for config.
* Update AWS plugin READMEs with credentials details, update Changelog.
* Fix changelog after rebasing to master and 0.13.1 release.
* Fix changelog after rebase.
Also changing the net_response and http_response plugins to only accept
duration strings for their timeout parameters. This is a breaking config
file change.
closes#1214
First is to write an internal CombinedOutput and Run function with a
timeout.
Second, the following instances of command runners need to have timeouts:
plugins/inputs/ping/ping.go
125: out, err := c.CombinedOutput()
plugins/inputs/exec/exec.go
91: if err := cmd.Run(); err != nil {
plugins/inputs/ipmi_sensor/command.go
31: err := cmd.Run()
plugins/inputs/sysstat/sysstat.go
194: out, err := cmd.CombinedOutput()
plugins/inputs/leofs/leofs.go
185: defer cmd.Wait()
plugins/inputs/sysstat/sysstat.go
282: if err := cmd.Wait(); err != nil {
closes#1067
this is so that we don't call os.Stat twice for every file matched
by Match(). Also changing the behavior to _not_ return the name of a
file that doesn't exist if it's not a glob.
changes:
- -sample-config will now comment out all but a few default plugins.
- config file parse errors will output path to bad conf file.
- cleanup 80-char line-length and some other style issues.
- default package conf file will now have all plugins, but commented
out.
closes#199closes#944
this includes:
- Add Accumulator to the Start() function of service inputs
- For message consumer plugins, use the Accumulator to constantly add
metrics and make Gather a dummy function
- rework unit tests to match this new behavior.
- make "flush_buffer_when_full" a config option that defaults to true
closes#666
This constitutes a large change in how we will parse different data
formats going forward (for the plugins that support it)
This is working off @henrypfhu's changes.
This will basically make the root directory a place for storing the
major telegraf interfaces, which will make telegraf's godoc looks quite
a bit nicer. And make it easier for contributors to lookup the few data
types that they actually care about.
closes#564
We are unifying the way that we handle configuration across the products
into the influxdata/config package. This provides the same API as
naoina/toml that was used previously, but provides some additional
features such as support for documenting generated TOML configs as well
as support for handling default options. This replaces all usage of
naoina/toml with influxdata/config.
Reuses same logic as the plugins for filtering points, should be only
a marginal performance decrease to check all the points before writing
to the output.
Added examples to the README as well (for generic pass/drop as well as
output pass/drop/tagpass/tagdrop).
X-Github-Closes #398closes#398closes#401
godep seems to have a problem when dependencies have `internal`
packages. So removing `godep go build` and `godep go test` from the
build process in favor of just checking out the correct revisions using
`godep restore` into the regular GOPATH.
This basically means that we are not actually using anything within the
Godeps directory except Godeps.json. I should probably make a separate
go dependency management system that does this.
Paul Myjavec
Piotr Popieluch
atzoum
Daniel Nelson
Maximilien Richer
DanKans
Seuf
Daniel Nelson
Patrick Hemmer
Daniel Nelson
tjmcs
Cameron Sparr
Alex Zorin
Jason Gardner
Mike Glazer
John Engelman
Pierre Fersing
Eugene Dementiev
Thibault Cohen
Sergio Jimenez
Henry Hu
Tim Raymond
Jack Zampolin
Philip Silva
Tait Clarridge