Skip to main content

Observability

KeyNetra includes first-party metrics and structured logging for operational visibility.

Observability components:

  • Metrics definitions: keynetra/observability/metrics.py
  • Metrics endpoint: keynetra/api/routes/metrics.py
  • Logging config: keynetra/infrastructure/logging.py
  • Request logging middleware: keynetra/api/middleware/logging.py

Metrics Endpoint​

GET /metrics returns Prometheus text format (text/plain; version=0.0.4).

Metric Families​

From implementation, key metrics include:

  • keynetra_access_checks_total
  • keynetra_acl_matches_total
  • keynetra_policy_evaluations_total
  • keynetra_relationship_traversals_total
  • keynetra_policy_compilations_total
  • keynetra_revision_updates_total
  • keynetra_access_check_latency_seconds
  • keynetra_decision_latency_seconds
  • keynetra_cache_hits_total
  • keynetra_cache_misses_total
  • keynetra_cache_events_total
  • keynetra_api_errors_total

These metrics cover authorization decisions, cache behavior, policy/model lifecycle, and API error rates.

Logging Modes​

  • JSON logs by default
  • Rich colored logs when KEYNETRA_LOG_FORMAT=rich

Docker startup script sets rich mode by default.

Use JSON mode for log aggregation pipelines and rich mode for local operator readability.

Prometheus and Grafana​

Compose stack includes monitoring:

  • Prometheus config: infra/docker/monitoring/prometheus/prometheus.yml
  • Grafana provisioning: infra/docker/monitoring/grafana/provisioning/
  • Dashboards: infra/docker/monitoring/grafana/dashboards/

Quick Validation​

curl -s http://localhost:8000/metrics | head