Traces and metrics are provided for observability of the orchestrator.
Traces provide paths of requests through the application. When tracing is enabled, traces an be viewed in an appropriate backend such as Jaeger, which provides a UI for viewing services and trace operations.
Metrics capture measurements at runtime of the application. When metrics are enabled, they can be viewed in an appropriate backend such as Prometheus, which provides a UI for exploring and graphing metrics.
Example server metrics:
incoming_request_count
success_response_count
server_error_response_count
Example orchestrator client metrics:
incoming_request_count
client_response_count
client_request_duration
Environment variables can be used to configure traces and/or metrics
- Use
OTLP_EXPORT
to provide one oftraces
,metrics
, or bothtraces,metrics
. - Use
OTEL_EXPORTER_OTLP_ENDPOINT
to configure an endpoint for traces or metrics e.g.http://collector-svc:4317
More details and configuration options are noted in the configuration section of the telemetry ADR.