correlation
Linking evidence from different signals about the same event, for example from a metric spike to a trace to its log lines.
In more detail
Correlation depends on shared identity: service name, route and status, the trace ID, and time. With those fields, an operator can move from a dashboard spike to the exact request that caused it in seconds.
Related terms
- trace ID — An identifier shared by every span and log line belonging to one request, so they can be found together.
- exemplar — A trace ID attached to one observation inside a histogram bucket, so a dashboard can jump from a slow bucket to an example request.
- resource attribute — An attribute describing the thing that produced telemetry, such as the service name, version or deployment environment.
- telemetry signal — One of the three main kinds of telemetry: metrics, logs or traces.
Taught in
- Bringing The Three Signals Together — Full-Stack Observability: Metrics, Tracing & Logging
- A Worked Example Postmortem — Full-Stack Observability: Metrics, Tracing & Logging