Glossary · Observability

What Are Exemplars?

Exemplars are the bridge from a metric spike to an actual trace that caused it, no guessing in between.

Exemplars are links from an aggregated metric, such as a specific histogram bucket, to an example trace that contributed to it, letting you jump directly from a metric spike to a concrete trace that illustrates it.

What they solve

Metrics tell you something changed but not which request; traces show individual requests but you have to find the right one. Exemplars close that gap by attaching, to a metric sample, the trace ID of a request that fell into that bucket, so a latency histogram's p99 bucket points straight at an example slow trace.

Why it matters for workflow

Without exemplars, moving from 'p99 latency spiked' to 'here is a slow request' means manual searching. With them it is one click from the metric to a representative trace, which collapses the time between noticing a symptom and seeing a concrete example of it.

How it fits Ops Singularity

Exemplar-style correlation, from an aggregate signal to a concrete trace, is exactly the linkage Sentinel AI follows automatically when it investigates a metric anomaly.

Frequently asked questions

What is an exemplar in metrics?

A reference attached to a metric sample that points to an example trace contributing to it, linking aggregate metrics to individual traces.

Why are exemplars useful?

They let you jump from a metric spike straight to a representative trace, instead of manually searching for a matching request.

One governed intelligence layer for every operation.

Ops Singularity turns open telemetry into autonomous, governed resolution. See it on your own stack.

Request a Demo → See TelemetryOps