Implement telegraf collecting stats on itself

closes #1348
This commit is contained in:
Cameron Sparr
2016-11-07 08:34:46 +00:00
parent d518d7d806
commit d71a42cd1b
26 changed files with 975 additions and 169 deletions

View File

@@ -7,6 +7,7 @@ import (
"github.com/influxdata/telegraf"
"github.com/influxdata/telegraf/internal/buffer"
"github.com/influxdata/telegraf/metric"
"github.com/influxdata/telegraf/selfstat"
)
const (
@@ -22,10 +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
}
@@ -51,7 +57,33 @@ 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
}
@@ -67,6 +99,7 @@ func (ro *RunningOutput) AddMetric(m telegraf.Metric) {
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.
@@ -85,28 +118,21 @@ func (ro *RunningOutput) AddMetric(m 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()
log.Printf("D! Output [%s] buffer fullness: %d / %d metrics. ",
ro.Name, nFails+nMetrics, ro.MetricBufferLimit)
ro.BufferSize.Incr(int64(nFails + nMetrics))
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
@@ -127,6 +153,7 @@ func (ro *RunningOutput) Write() error {
if err == nil {
err = ro.write(batch)
}
if err != nil {
ro.failMetrics.Add(batch...)
return err
@@ -135,17 +162,19 @@ 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.BufferSize.Incr(-int64(nMetrics))
ro.WriteTime.Incr(elapsed.Nanoseconds())
}
return err
}