Observability

Production systems need visibility beyond log lines. mangoo I/O exposes request metrics (counts and latency) for the admin dashboard and optional OpenTelemetry export for traces. Both are off by default so local development stays lightweight; enable them explicitly when you deploy.

Metrics answer "how much traffic and how slow?" from inside the app itself. OTLP instead sends spans out to a collector (Jaeger, Grafana Tempo, vendor backends) for distributed tracing across services. See Operating for JVM and vault setup alongside these features.

Metrics

Set metrics.enable to true to start counting requests and recording processing time. Values show up on the admin dashboard. Metrics stay off by default so they never add overhead you did not ask for.

metrics:
  enable: true

Admin UI traffic is excluded from the counters.

OpenTelemetry

mangoo I/O can export traces over OTLP/gRPC:

otlp:
  enable: true
  endpoint: http://127.0.0.1:4317

When enabled, spans go out under the service name io.mangoo along with the current application mode. Leave otlp.enable false unless you actually run a collector to receive them.

There is no public /health HTTP endpoint. If you need a MongoDB ping for your own health checks, call Datastore.isHealthy() directly.