diff --git a/CHANGELOG.md b/CHANGELOG.md index 5137b86df..6a862a0db 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -3,6 +3,7 @@ ### Features - [#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. ### Bugfixes diff --git a/etc/telegraf.conf b/etc/telegraf.conf index 10e949302..c667c4c9f 100644 --- a/etc/telegraf.conf +++ b/etc/telegraf.conf @@ -666,6 +666,11 @@ # 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 # # Read statistics from one or many dovecot servers diff --git a/plugins/inputs/docker/docker.go b/plugins/inputs/docker/docker.go index dfd768c1a..e3876bd64 100644 --- a/plugins/inputs/docker/docker.go +++ b/plugins/inputs/docker/docker.go @@ -25,6 +25,8 @@ type Docker struct { Endpoint string ContainerNames []string Timeout internal.Duration + PerDevice bool `toml:"perdevice"` + Total bool `toml:"total"` client DockerClient } @@ -58,6 +60,13 @@ var sampleConfig = ` 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 + ` // Description returns input description @@ -246,7 +255,7 @@ func (d *Docker) gatherContainer( tags[k] = label } - gatherContainerStats(v, acc, tags, container.ID) + gatherContainerStats(v, acc, tags, container.ID, d.PerDevice, d.Total) return nil } @@ -256,6 +265,8 @@ func gatherContainerStats( acc telegraf.Accumulator, tags map[string]string, id string, + perDevice bool, + total bool, ) { now := stat.Read @@ -323,6 +334,7 @@ func gatherContainerStats( acc.AddFields("docker_container_cpu", fields, percputags, now) } + totalNetworkStatMap := make(map[string]interface{}) for network, netstats := range stat.Networks { netfields := map[string]interface{}{ "rx_dropped": netstats.RxDropped, @@ -336,12 +348,35 @@ func gatherContainerStats( "container_id": id, } // Create a new network tag dictionary for the "network" tag - nettags := copyTags(tags) - nettags["network"] = network - acc.AddFields("docker_container_net", netfields, nettags, now) + if perDevice { + nettags := copyTags(tags) + nettags["network"] = network + acc.AddFields("docker_container_net", netfields, nettags, now) + } + if total { + for field, value := range netfields { + if field == "container_id" { + continue + } + _, ok := totalNetworkStatMap[field] + if ok { + totalNetworkStatMap[field] = totalNetworkStatMap[field].(uint64) + value.(uint64) + } else { + totalNetworkStatMap[field] = value + } + } + } } - gatherBlockIOMetrics(stat, acc, tags, now, id) + // totalNetworkStatMap could be empty if container is running with --net=host. + if total && len(totalNetworkStatMap) != 0 { + nettags := copyTags(tags) + nettags["network"] = "total" + totalNetworkStatMap["container_id"] = id + acc.AddFields("docker_container_net", totalNetworkStatMap, nettags, now) + } + + gatherBlockIOMetrics(stat, acc, tags, now, id, perDevice, total) } func calculateMemPercent(stat *types.StatsJSON) float64 { @@ -370,6 +405,8 @@ func gatherBlockIOMetrics( tags map[string]string, now time.Time, id string, + perDevice bool, + total bool, ) { blkioStats := stat.BlkioStats // Make a map of devices to their block io stats @@ -431,11 +468,33 @@ func gatherBlockIOMetrics( deviceStatMap[device]["sectors_recursive"] = metric.Value } + totalStatMap := make(map[string]interface{}) for device, fields := range deviceStatMap { - iotags := copyTags(tags) - iotags["device"] = device fields["container_id"] = id - acc.AddFields("docker_container_blkio", fields, iotags, now) + if perDevice { + iotags := copyTags(tags) + iotags["device"] = device + acc.AddFields("docker_container_blkio", fields, iotags, now) + } + if total { + for field, value := range fields { + if field == "container_id" { + continue + } + _, ok := totalStatMap[field] + if ok { + totalStatMap[field] = totalStatMap[field].(uint64) + value.(uint64) + } else { + totalStatMap[field] = value + } + } + } + } + if total { + totalStatMap["container_id"] = id + iotags := copyTags(tags) + iotags["device"] = "total" + acc.AddFields("docker_container_blkio", totalStatMap, iotags, now) } } @@ -480,7 +539,8 @@ func parseSize(sizeStr string) (int64, error) { func init() { inputs.Add("docker", func() telegraf.Input { return &Docker{ - Timeout: internal.Duration{Duration: time.Second * 5}, + PerDevice: true, + Timeout: internal.Duration{Duration: time.Second * 5}, } }) } diff --git a/plugins/inputs/docker/docker_test.go b/plugins/inputs/docker/docker_test.go index b1c76f5af..9f2e97f73 100644 --- a/plugins/inputs/docker/docker_test.go +++ b/plugins/inputs/docker/docker_test.go @@ -24,7 +24,7 @@ func TestDockerGatherContainerStats(t *testing.T) { "container_name": "redis", "container_image": "redis/image", } - gatherContainerStats(stats, &acc, tags, "123456789") + gatherContainerStats(stats, &acc, tags, "123456789", true, true) // test docker_container_net measurement netfields := map[string]interface{}{ @@ -42,6 +42,21 @@ func TestDockerGatherContainerStats(t *testing.T) { nettags["network"] = "eth0" acc.AssertContainsTaggedFields(t, "docker_container_net", netfields, nettags) + netfields = map[string]interface{}{ + "rx_dropped": uint64(6), + "rx_bytes": uint64(8), + "rx_errors": uint64(10), + "tx_packets": uint64(12), + "tx_dropped": uint64(6), + "rx_packets": uint64(8), + "tx_errors": uint64(10), + "tx_bytes": uint64(12), + "container_id": "123456789", + } + nettags = copyTags(tags) + nettags["network"] = "total" + acc.AssertContainsTaggedFields(t, "docker_container_net", netfields, nettags) + // test docker_blkio measurement blkiotags := copyTags(tags) blkiotags["device"] = "6:0" @@ -52,6 +67,15 @@ func TestDockerGatherContainerStats(t *testing.T) { } acc.AssertContainsTaggedFields(t, "docker_container_blkio", blkiofields, blkiotags) + blkiotags = copyTags(tags) + blkiotags["device"] = "total" + blkiofields = map[string]interface{}{ + "io_service_bytes_recursive_read": uint64(100), + "io_serviced_recursive_write": uint64(302), + "container_id": "123456789", + } + acc.AssertContainsTaggedFields(t, "docker_container_blkio", blkiofields, blkiotags) + // test docker_container_mem measurement memfields := map[string]interface{}{ "max_usage": uint64(1001), @@ -186,6 +210,17 @@ func testStats() *types.StatsJSON { TxBytes: 4, } + stats.Networks["eth1"] = types.NetworkStats{ + RxDropped: 5, + RxBytes: 6, + RxErrors: 7, + TxPackets: 8, + TxDropped: 5, + RxPackets: 6, + TxErrors: 7, + TxBytes: 8, + } + sbr := types.BlkioStatEntry{ Major: 6, Minor: 0, @@ -198,11 +233,19 @@ func testStats() *types.StatsJSON { Op: "write", Value: 101, } + sr2 := types.BlkioStatEntry{ + Major: 6, + Minor: 1, + Op: "write", + Value: 201, + } stats.BlkioStats.IoServiceBytesRecursive = append( stats.BlkioStats.IoServiceBytesRecursive, sbr) stats.BlkioStats.IoServicedRecursive = append( stats.BlkioStats.IoServicedRecursive, sr) + stats.BlkioStats.IoServicedRecursive = append( + stats.BlkioStats.IoServicedRecursive, sr2) return stats }