-
Notifications
You must be signed in to change notification settings - Fork 111
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Add alerts for internal metrics (#1412)
- Loading branch information
Showing
7 changed files
with
117 additions
and
0 deletions.
There are no files selected for viewing
This file contains 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 |
|---|---|---|
| @@ -0,0 +1,3 @@ | ||
| # Beyla monitoring mixins | ||
|
|
||
| This mixin includes Grafana Dashboards and Alerts for Beyla. |
This file contains 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 |
|---|---|---|
| @@ -0,0 +1,97 @@ | ||
| { | ||
| prometheusAlerts+: { | ||
| groups+: [ | ||
| { | ||
| name: 'beyla_internal_alerts', | ||
| rules: [ | ||
| { | ||
| alert: 'BeylaOTELMetricExportHighErrorRate', | ||
| expr: ||| | ||
| 100 * sum(rate(beyla_otel_metric_export_errors_total[1m])) by (cluster) | ||
| / | ||
| sum(rate(beyla_otel_metric_exports_total[1m])) by (cluster) | ||
| > 2 | ||
| |||, | ||
| 'for': '15m', | ||
| labels: { | ||
| severity: 'warning', | ||
| }, | ||
| annotations: { | ||
| summary: 'Beyla has a high error rate for OTEL metric exports', | ||
| description: 'Beyla in { $labels.cluster }} is experiencing {{ printf "%.2f" $value }}% errors for OTEL metric exports.', | ||
| }, | ||
| }, | ||
| { | ||
| alert: 'BeylaOTELMetricExportHighErrorRate', | ||
| expr: ||| | ||
| 100 * sum(rate(beyla_otel_metric_export_errors_total[1m])) by (cluster) | ||
| / | ||
| sum(rate(beyla_otel_metric_exports_total[1m])) by (cluster) | ||
| > 5 | ||
| |||, | ||
| 'for': '15m', | ||
| labels: { | ||
| severity: 'critical', | ||
| }, | ||
| annotations: { | ||
| summary: 'Beyla has a high error rate for OTEL metric exports', | ||
| description: 'Beyla in { $labels.cluster }} is experiencing {{ printf "%.2f" $value }}% errors for OTEL metric exports.', | ||
| }, | ||
| }, | ||
| { | ||
| alert: 'BeylaOTELTraceExportHighErrorRate', | ||
| expr: ||| | ||
| 100 * sum(rate(beyla_otel_trace_export_errors_total[1m])) by (cluster) | ||
| / | ||
| sum(rate(beyla_otel_trace_exports_total[1m])) by (cluster) | ||
| > 2 | ||
| |||, | ||
| 'for': '15m', | ||
| labels: { | ||
| severity: 'warning', | ||
| }, | ||
| annotations: { | ||
| summary: 'Beyla has a high error rate for OTEL trace exports', | ||
| description: 'Beyla in { $labels.cluster }} is experiencing {{ printf "%.2f" $value }}% errors for OTEL trace exports.', | ||
| }, | ||
| }, | ||
| { | ||
| alert: 'BeylaOTELTraceExportHighErrorRate', | ||
| expr: ||| | ||
| 100 * sum(rate(beyla_otel_trace_export_errors_total[1m])) by (cluster) | ||
| / | ||
| sum(rate(beyla_otel_trace_exports_total[1m])) by (cluster) | ||
| > 5 | ||
| |||, | ||
| 'for': '15m', | ||
| labels: { | ||
| severity: 'critical', | ||
| }, | ||
| annotations: { | ||
| summary: 'Beyla has a high error rate for OTEL trace exports', | ||
| description: 'Beyla in { $labels.cluster }} is experiencing {{ printf "%.2f" $value }}% errors for OTEL trace exports.', | ||
| }, | ||
| }, | ||
| { | ||
| alert: 'BeylaInstrumentedProcessesNoTelemetry', | ||
| expr: ||| | ||
| sum by(cluster) (beyla_instrumented_processes{process_name!="beyla"}) > 1 and | ||
| (absent(sum by(cluster) (rate(beyla_otel_metric_exports_total[5m]))) or | ||
| absent(sum by(cluster) (rate(beyla_otel_trace_exports_total[5m]))) or | ||
| absent(sum by(cluster) (rate(beyla_ebpf_tracer_flushes_sum[5m]))) or | ||
| absent(sum by(cluster) (rate(beyla_prometheus_http_requests_total[5m])))) | ||
| |||, | ||
| 'for': '10m', | ||
| labels: { | ||
| severity: 'warning', | ||
| }, | ||
| annotations: { | ||
| summary: 'Beyla has instrumented processes without metrics or traces', | ||
| description: 'Beyla in { $labels.cluster }} has are more than 1 instrumented processes, but no metrics or traces have been exported in the last 10 minutes.', | ||
| }, | ||
| }, | ||
| ], | ||
| }, | ||
| ], | ||
| }, | ||
| } |
File renamed without changes.
File renamed without changes.
File renamed without changes.
This file contains 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 |
|---|---|---|
| @@ -0,0 +1,15 @@ | ||
| { | ||
| grafanaDashboardFolder+: 'beyla', | ||
| grafanaDashboards+: { | ||
| 'beyla_debug.json': (import './beyla_debug.json') { | ||
| uid: std.md5('beyla_debug.json'), | ||
| }, | ||
| 'application.json': (import './application.json') { | ||
| uid: std.md5('application.json'), | ||
| }, | ||
| // TODO: Include this dashboard only if application_process is enabled | ||
| 'application_process.json': (import './application_process.json') { | ||
| uid: std.md5('application_process.json'), | ||
| }, | ||
| }, | ||
| } |
This file contains 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 |
|---|---|---|
| @@ -0,0 +1,2 @@ | ||
| (import './alerts/alerts.libsonnet') | ||
| + (import './dashboards/dashboards.libsonnet') |