Create a custom dashboard when a team needs selected trace-derived metrics alongside other operational signals. The available widget names and editing controls depend on the dashboard capabilities enabled for your deployment.

Build the dashboard

1. Define the audience and decision the dashboard must support.
2. Create or open a dashboard and add a metric visualization.
3. Select trace-derived latency, error-rate, throughput, or Apdex metrics that are available in the metric catalog.
4. Filter with stable labels such as service, operation, and environment.
5. Choose an aggregation and interval appropriate for the traffic volume.
6. Add clear titles, units, thresholds, and a consistent time range.
7. Save and verify the dashboard against the trace service view.

Avoid one widget per ephemeral instance or unbounded attribute value. Start with service-level health, then add operation-level detail only when it supports an action. A custom dashboard does not change metric generation or recover traces removed by sampling.