Skip to main content

Prometheus Metrics Reference

CronJob Guardian exports Prometheus metrics on port 8080 at /metrics.

CronJob Metrics​

cronjob_guardian_success_rate​

Current success rate percentage for each monitored CronJob.

LabelDescription
namespaceCronJob namespace
cronjobCronJob name
monitorMonitor name

Type: Gauge

Example:

cronjob_guardian_success_rate{namespace="production", cronjob="daily-backup"}

cronjob_guardian_duration_seconds​

Execution duration histogram.

LabelDescription
namespaceCronJob namespace
cronjobCronJob name

Type: Histogram

Buckets: 1s, 5s, 10s, 30s, 60s, 300s, 600s, 1800s, 3600s

Example:

# P95 duration
histogram_quantile(0.95, sum(rate(cronjob_guardian_duration_seconds_bucket[5m])) by (le, cronjob))

cronjob_guardian_executions_total​

Total number of job executions.

LabelDescription
namespaceCronJob namespace
cronjobCronJob name
statusExecution status (success, failed)

Type: Counter

Example:

# Failure rate
rate(cronjob_guardian_executions_total{status="failed"}[1h])

cronjob_guardian_active_alerts​

Number of currently active alerts.

LabelDescription
namespaceCronJob namespace
cronjobCronJob name
severityAlert severity

Type: Gauge

Example:

cronjob_guardian_active_alerts{severity="critical"} > 0

Alert Metrics​

cronjob_guardian_alerts_total​

Total alerts sent.

LabelDescription
typeAlert type (failure, deadManSwitch, slaViolation, durationRegression)
severityAlert severity (critical, warning, info)
channelChannel name

Type: Counter

Example:

# Alert rate by channel
sum(rate(cronjob_guardian_alerts_total[1h])) by (channel)

cronjob_guardian_alert_dispatch_duration_seconds​

Time to dispatch alerts to channels.

LabelDescription
channelChannel name
successWhether dispatch succeeded

Type: Histogram

Operator Metrics​

cronjob_guardian_leader_status​

Whether this instance is the leader (1) or standby (0).

Type: Gauge

Example:

# Ensure a leader exists
sum(cronjob_guardian_leader_status) == 1

cronjob_guardian_db_connections_open​

Number of open database connections.

Type: Gauge

cronjob_guardian_db_connections_idle​

Number of idle database connections.

Type: Gauge

cronjob_guardian_reconcile_total​

Total reconciliation operations.

LabelDescription
controllerController name
resultResult (success, error, requeue)

Type: Counter

cronjob_guardian_reconcile_duration_seconds​

Reconciliation duration.

LabelDescription
controllerController name

Type: Histogram

Standard Metrics​

CronJob Guardian also exports standard controller-runtime metrics:

  • controller_runtime_reconcile_total
  • controller_runtime_reconcile_errors_total
  • controller_runtime_reconcile_time_seconds
  • workqueue_* metrics

And standard Go runtime metrics:

  • go_goroutines
  • go_memstats_*
  • process_*

Example Queries​

CronJob Health​

# CronJobs with success rate below 95%
cronjob_guardian_success_rate < 95

# Slowest CronJobs by P95 duration
topk(10, histogram_quantile(0.95,
  sum(rate(cronjob_guardian_duration_seconds_bucket[1h])) by (le, namespace, cronjob)
))

# CronJobs with active critical alerts
cronjob_guardian_active_alerts{severity="critical"} > 0

Alert Analysis​

# Total alerts in last 24 hours
sum(increase(cronjob_guardian_alerts_total[24h]))

# Most alerting CronJobs
topk(10, sum(increase(cronjob_guardian_alerts_total[24h])) by (namespace, cronjob))

# Channel alert rate
sum(rate(cronjob_guardian_alerts_total[1h])) by (channel)

Operator Health​

# Operator up
up{job="cronjob-guardian"}

# Leader elected
cronjob_guardian_leader_status == 1

# Reconciliation error rate
sum(rate(cronjob_guardian_reconcile_total{result="error"}[5m]))

Grafana Dashboard​

See Prometheus Integration for a complete Grafana dashboard configuration.