feat: add configurable metric export interval to lambda telemetry API receiver #2114
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -81,6 +81,9 @@ type telemetryAPIReceiver struct { | |
| faaSMetricBuilders *FaaSMetricBuilders | ||
| currentFaasInvocationID string | ||
| logReport bool | ||
| exportInterval time.Duration | ||
| stopCh chan struct{} | ||
| wg sync.WaitGroup | ||
| } | ||
|
|
||
| func (r *telemetryAPIReceiver) Start(ctx context.Context, host component.Host) error { | ||
|
|
@@ -94,6 +97,11 @@ func (r *telemetryAPIReceiver) Start(ctx context.Context, host component.Host) e | |
| _ = r.httpServer.ListenAndServe() | ||
| }() | ||
|
|
||
| if r.exportInterval > 0 { | ||
| r.wg.Add(1) | ||
| go r.startMetricsExporter() | ||
| } | ||
|
|
||
| telemetryClient := telemetryapi.NewClient(r.logger) | ||
| if len(r.types) > 0 { | ||
| _, err := telemetryClient.Subscribe(ctx, r.types, r.extensionID, fmt.Sprintf("http://%s/", address)) | ||
|
|
@@ -106,9 +114,60 @@ func (r *telemetryAPIReceiver) Start(ctx context.Context, host component.Host) e | |
| } | ||
|
|
||
| func (r *telemetryAPIReceiver) Shutdown(ctx context.Context) error { | ||
| close(r.stopCh) | ||
| r.wg.Wait() | ||
| if r.exportInterval > 0 { | ||
| r.flushMetrics(ctx) | ||
| } | ||
| return nil | ||
| } | ||
|
|
||
| func (r *telemetryAPIReceiver) startMetricsExporter() { | ||
| defer r.wg.Done() | ||
| ticker := time.NewTicker(r.exportInterval) | ||
| defer ticker.Stop() | ||
|
|
||
| for { | ||
| select { | ||
| case <-ticker.C: | ||
| r.flushMetrics(context.Background()) | ||
| case <-r.stopCh: | ||
| return | ||
| } | ||
| } | ||
| } | ||
|
|
||
| func (r *telemetryAPIReceiver) flushMetrics(ctx context.Context) { | ||
| r.mu.Lock() | ||
| defer r.mu.Unlock() | ||
| r.flushMetricsLocked(ctx) | ||
| } | ||
|
|
||
| func (r *telemetryAPIReceiver) flushMetricsLocked(ctx context.Context) { | ||
| metric := pmetric.NewMetrics() | ||
| resourceMetric := metric.ResourceMetrics().AppendEmpty() | ||
| r.resource.CopyTo(resourceMetric.Resource()) | ||
| scopeMetric := resourceMetric.ScopeMetrics().AppendEmpty() | ||
| scopeMetric.Scope().SetName(scopeName) | ||
| scopeMetric.SetSchemaUrl(semconv.SchemaURL) | ||
|
|
||
| ts := pcommon.NewTimestampFromTime(time.Now()) | ||
| r.faaSMetricBuilders.coldstartsMetric.AppendDataPoints(scopeMetric, ts) | ||
| r.faaSMetricBuilders.errorsMetric.AppendDataPoints(scopeMetric, ts) | ||
| r.faaSMetricBuilders.timeoutsMetric.AppendDataPoints(scopeMetric, ts) | ||
| r.faaSMetricBuilders.initDurationMetric.AppendDataPoints(scopeMetric, ts) | ||
| r.faaSMetricBuilders.memUsageMetric.AppendDataPoints(scopeMetric, ts) | ||
| r.faaSMetricBuilders.invocationsMetric.AppendDataPoints(scopeMetric, ts) | ||
| r.faaSMetricBuilders.invokeDurationMetric.AppendDataPoints(scopeMetric, ts) | ||
|
|
||
| if metric.MetricCount() > 0 { | ||
|
Member
There was a problem hiding this comment. Edge case here but think of the following scenario:
receivers:
telemetryapi:
types:
- platform
metrics_temporality: "delta"
...
service:
pipelines:
traces:
receivers: [telemetryapi]
exporters: [otlp]This leds to Whenever the ticker fires and I believe that is problematic, because now following calls inside that same So the check for
Member
There was a problem hiding this comment. Also, even when properly using an actual metrics pipeline We run the risk of just producing metrics with 0 datapoints when temporality is delta, which seems useless / wasteful. |
||
| err := r.nextMetrics.ConsumeMetrics(ctx, metric) | ||
| if err != nil { | ||
| r.logger.Error("error flushing metrics", zap.Error(err)) | ||
| } | ||
| } | ||
| } | ||
|
|
||
| func newSpanID() pcommon.SpanID { | ||
| sid := pcommon.SpanID{} | ||
| _, _ = crand.Read(sid[:]) | ||
|
|
@@ -198,13 +257,9 @@ func (r *telemetryAPIReceiver) httpHandler(w http.ResponseWriter, req *http.Requ | |
| } | ||
| // Metrics | ||
| if r.nextMetrics != nil { | ||
| r.recordMetrics(slice) | ||
| if r.exportInterval == 0 { | ||
| r.flushMetricsLocked(context.Background()) | ||
| } | ||
| } | ||
|
|
||
|
|
@@ -233,37 +288,22 @@ func (r *telemetryAPIReceiver) getRecordRequestId(record map[string]interface{}) | |
| return "" | ||
| } | ||
|
|
||
| func (r *telemetryAPIReceiver) recordMetrics(slice []event) { | ||
| for _, el := range slice { | ||
| record, ok := el.Record.(map[string]any) | ||
| if !ok { | ||
| continue | ||
| } | ||
|
|
||
| switch el.Type { | ||
| case string(telemetryapi.PlatformInitStart): | ||
| r.faaSMetricBuilders.coldstartsMetric.Add(1) | ||
| case string(telemetryapi.PlatformInitReport): | ||
| status, _ := record["status"].(string) | ||
| if status == telemetryFailureStatus || status == telemetryErrorStatus { | ||
| r.faaSMetricBuilders.errorsMetric.Add(1) | ||
| } else if status == telemetryTimeoutStatus { | ||
| r.faaSMetricBuilders.timeoutsMetric.Add(1) | ||
| } | ||
|
|
||
| metrics, ok := record["metrics"].(map[string]any) | ||
|
|
@@ -277,7 +317,6 @@ func (r *telemetryAPIReceiver) createMetrics(slice []event) (pmetric.Metrics, er | |
| } | ||
|
|
||
| r.faaSMetricBuilders.initDurationMetric.Record(durationMs / 1000.0) | ||
| case string(telemetryapi.PlatformReport): | ||
| metrics, ok := record["metrics"].(map[string]any) | ||
| if !ok { | ||
|
|
@@ -287,20 +326,16 @@ func (r *telemetryAPIReceiver) createMetrics(slice []event) (pmetric.Metrics, er | |
| maxMemoryUsedMb, ok := metrics["maxMemoryUsedMB"].(float64) | ||
| if ok { | ||
| r.faaSMetricBuilders.memUsageMetric.Record(maxMemoryUsedMb * 1000000.0) | ||
| } | ||
| case string(telemetryapi.PlatformRuntimeDone): | ||
| status, _ := record["status"].(string) | ||
|
|
||
| if status == telemetrySuccessStatus { | ||
| r.faaSMetricBuilders.invocationsMetric.Add(1) | ||
| } else if status == telemetryFailureStatus || status == telemetryErrorStatus { | ||
| r.faaSMetricBuilders.errorsMetric.Add(1) | ||
| } else if status == telemetryTimeoutStatus { | ||
| r.faaSMetricBuilders.timeoutsMetric.Add(1) | ||
| } | ||
|
|
||
| metrics, ok := record["metrics"].(map[string]any) | ||
|
|
@@ -311,11 +346,9 @@ func (r *telemetryAPIReceiver) createMetrics(slice []event) (pmetric.Metrics, er | |
| durationMs, ok := metrics["durationMs"].(float64) | ||
| if ok { | ||
| r.faaSMetricBuilders.invokeDurationMetric.Record(durationMs / 1000.0) | ||
| } | ||
| } | ||
| } | ||
| } | ||
|
|
||
| func (r *telemetryAPIReceiver) createLogs(slice []event) (plog.Logs, error) { | ||
|
|
@@ -675,6 +708,8 @@ func newTelemetryAPIReceiver( | |
| resource: r, | ||
| faaSMetricBuilders: NewFaaSMetricBuilders(pcommon.NewTimestampFromTime(time.Now()), getMetricsTemporality(cfg)), | ||
| logReport: cfg.LogReport, | ||
| exportInterval: time.Duration(cfg.ExportInterval) * time.Millisecond, | ||
| stopCh: make(chan struct{}), | ||
| }, nil | ||
| } | ||
|
|
||
|
|
||
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should we be shutting down the httpServer here first?