Monitoring and alerts

Metrics endpoint

• Prometheus metrics: GET /actuator/prometheus
• Health probes:
  • GET /actuator/health/liveness
  • GET /actuator/health/readiness

Configuration

• DWARVENPICK_METRICS_PROMETHEUS_ENABLED (default: true) enables the Prometheus endpoint.
• Helm chart value: metrics.prometheus.enabled

Kubernetes scraping

The Helm chart exposes the backend on .Values.service.port (default 8080). Configure your Prometheus instance to scrape:

• http://<service-name>:8080/actuator/prometheus

If you use Prometheus Operator, create a ServiceMonitor that targets the backend Service and port.

Key metrics

• Query lifecycle:
  • dwarvenpick_query_active{status="queued|running"}
  • dwarvenpick_query_execution_total{outcome=...}
  • dwarvenpick_query_duration_seconds{outcome=...}
  • dwarvenpick_query_cancel_total
  • dwarvenpick_query_timeout_total
• Exports:
  • dwarvenpick_query_export_attempts_total{outcome=...}
• Auth:
  • dwarvenpick_auth_login_attempts_total{provider=...,outcome=...}
• Pools:
  • dwarvenpick_pool_active
  • dwarvenpick_pool_idle
  • dwarvenpick_pool_total

Recommended alerts

1. High query failure rate:
   • Trigger: increase(dwarvenpick_query_execution_total{outcome="failed"}[5m]) / increase(dwarvenpick_query_execution_total[5m]) > 0.2
2. Query timeout burst:
   • Trigger: increase(dwarvenpick_query_timeout_total[5m]) > 10
3. High queue pressure:
   • Trigger: dwarvenpick_query_active{status="queued"} > 20 for 10m
4. Pool saturation:
   • Trigger: dwarvenpick_pool_active / dwarvenpick_pool_total > 0.9 for 5m
5. Login failure surge:
   • Trigger: increase(dwarvenpick_auth_login_attempts_total{outcome="failed"}[5m]) > 25