Observability

Wow exposes two different views of the same runtime pipeline:

• Metrics aggregate finite operations, receiver streams, and batch activity with bounded Micrometer tags.
• wow-opentelemetry creates OpenTelemetry spans and propagates message trace context across command, event, state, persistence, and wait boundaries.

Use metrics to locate a failing or slow stage, then use traces to follow one execution. Neither signal replaces backend reconciliation, projection-lag checks, or deployment evidence.

Where Each Signal Is Documented

Signal	Page
Metrics: names, tags, storage batch and query metrics, export	Metrics
Traces: installation, instrumented stages, span names and attributes, your own spans	OpenTelemetry
Switches and exporter wiring	Observability Configuration

Investigate an Incident

For one incident, follow this evidence chain:

1. select the metric stage and time window by component, operation, context, aggregate, and outcome;
2. find a matching trace by service, operation, aggregate, request ID, or message ID;
3. confirm the span reaches the expected store and downstream processor;
4. reconcile the final backend version/read model and deployment revision separately.