Observability Practical By Samson Tanimawo, PhD Published Jul 8, 2026 4 min read

The Exemplar Pattern: Metrics to Traces

Exemplars link a slow metric data point to the trace that produced it. The pattern, the OTel support, and what teams gain.

The idea

Each metric data point can carry an exemplar: a sample trace ID that contributed to that data point.

Click a slow p99 latency point on a dashboard; jump to the actual trace. The debugging loop closes.

OpenTelemetry SDKs support exemplars. Storage backends (Prometheus, Cortex) support them. Vendors are catching up.

Implementation is mostly configuration; the cost is small.

Mean-time-to-context drops dramatically. Engineers debug from data, not from guesses.

Particularly valuable for tail-latency investigations. The exemplar is the slow trace by definition.