Use trace context and shared resource attributes to move from a slow or failed request to related logs, metrics, and infrastructure resources. Correlation narrows the investigation without assuming that every telemetry signal has the same retention, sampling, or access policy.
Start with a trace or service and a precise time range. Record the trace ID, service, operation, resource attributes, and failure time, then use those values to inspect the other telemetry views.Correlation workflows
- Traces and Logs explains how trace identifiers and service context connect request flow to log records.
- Traces and Metrics uses latency, error rate, throughput, and resource metrics to establish scope and timing.
- Traces and Infrastructure Resources maps traced services to Kubernetes and other resource context carried on spans.
Keep the evidence aligned
Use the same client, time zone, and time window in each view. Sampling can remove individual traces even when independently generated metrics still show an aggregate symptom. A missing correlated result therefore does not prove that the event did not occur; verify retention, permissions, attribute presence, and clock alignment before concluding that telemetry is absent.