StatsD configuration (default)
StatsD resource limits are managed at the API level viacomponentsConfig and apply to all components — they can’t be configured independently per component.
Airflow configuration
Prometheus integration
Grafana dashboards
Access Grafana at:- Airflow Dag performance
- Task execution metrics
- Scheduler health
- Worker utilization
Alerting
Built-in alerts
AirflowDeploymentUnhealthyAirflowSchedulerUnhealthyAirflowTasksPendingIncreasing
Key metrics
Best practices
- Set appropriate retention based on storage capacity.
- Use OTEL for multi-backend export.
- Configure alerts for critical health metrics.
- Monitor task queue depth for scaling needs.