Metrics
The orchestrator exposes Prometheus-compatible metrics on a dedicated HTTP endpoint. These cover workflow execution, task performance, queue depth, database pool health, journal cache efficiency, polling latency, retry rates, and dead letter queue depth.
- Prometheus-native — Standard
/metricsendpoint with all metrics following Prometheus naming conventions - Dedicated metrics port — Metrics are served on port 8001, separate from the gRPC (50051) and HTTP (8080) APIs
- Comprehensive coverage — 100+ metrics across workflows, tasks, queues, storage, polling, retries, and DLQ
- Zero-config — Sensible defaults with environment variable overrides
Metrics Endpoint
The orchestrator serves metrics at:
1http://<host>:8001/metricsA health endpoint is also available:
1http://<host>:8001/healthConfiguration
1# Enable metrics (enabled by default)2ORCHER_METRICS_ENABLED=true34# Metrics bind address5ORCHER_METRICS_ADDRESS="0.0.0.0:8001"67# Log level and format8LOG_LEVEL=info9LOG_FORMAT=json # json or prettyNext Steps
- Metric Reference — Full list of all exposed metrics
- Prometheus Setup — Scrape configuration and useful PromQL queries
- Alerting — Recommended alert rules
- Logging — Structured logging configuration
- Tracing — Distributed tracing with OpenTelemetry
- Debugging — Debugging workflow execution
- Storage — PostgreSQL and Redis configuration