Question

A microservice running on GKE emits structured JSON logs to Cloud Logging and is instrumented with OpenTelemetry to export spans to Cloud Trace. During an incident, an SRE opens a slow trace in Cloud Trace but cannot jump from a specific span to the application log entries produced while handling that request. The team wants clicking a trace to reveal the correlated logs (and vice versa) without changing the sampling strategy. What is the most effective way to enable this correlation?

Accepted Answer

Populate the LogEntry 'trace' field with projects/[PROJECT_ID]/traces/[TRACE_ID] and set the 'spanId' field so Cloud Logging associates each log entry with its trace and span.

Answer

Increase the Cloud Trace sampling rate to 100% so every request generates a span that Cloud Logging can automatically link to logs.

Answer

Create a log-based metric keyed on the request latency and add it as a chart to the Cloud Trace analysis view.

Answer

Export both logs and traces to the same BigQuery dataset and run a JOIN query on the timestamp columns to reconstruct the relationship.