Service detail and the service map

Service detail and the service map

Open a service to see its dashboard: latency percentiles (p50/p95/p99), error rate, throughput, top endpoints, and its place in the service map (dependencies with health and latency on each edge).

Cookbook: find why a service is slow

1. Open the service; note the p95 latency trend and any error spike.
2. Look at top endpoints — sort by latency to find the worst operation.
3. On the service map, check the edges to downstream dependencies — a red/slow edge points at the culprit (a database, a payment provider).
4. Click into traces for that endpoint to open the waterfall (next section).

Use cases

• Deploy regression: compare the service's latency before/after a release time.
• Noisy neighbor: spot a dependency dragging several services down on the map.
• Error surge: jump from the error-rate spike straight to the failing traces.