Skip to main content
The Services workspace turns incoming OpenTelemetry traces into a service-focused investigation view. Use it to answer which services are reporting, receiving the most traffic, returning errors, or showing high latency.

​
Open the service catalog

Choose Services under Observe. The catalog reports measurements from incoming server and consumer spans in the selected time window:
  • Request and error counts
  • Error rate
  • Approximate p50 and p95 latency
  • Last-seen time
Search by service name, filter by deployment environment, change the time window, or sort by traffic, error rate, p95 latency, or recency.

​
Investigate a service

Select a service to open its detail view. The active tab is encoded in the URL, so the investigation can be bookmarked or shared.
  • Overview groups incoming operations and compares request count, error rate, and latency.
  • HTTP groups server spans by method and route.
  • Database groups database calls by system and operation.
  • Queues groups messaging activity by system, destination, and operation type.
  • Errors shows only failing incoming operations.
Select an operation or route to open matching traces without changing the active time window. The service header also links to matching traces, logs, metrics, the service map, and AI-assisted investigation.

​
Exact time windows

Relative windows such as 15 minutes, 1 hour, 24 hours, and 7 days are available from the toolbar. When Services is opened from an investigation with exact from and to timestamps, those timestamps remain attached to service filters and drilldowns so every view examines the same evidence.

​
Telemetry requirements

Services requires OpenTelemetry traces with a non-empty service name and incoming SERVER or CONSUMER span kinds. HTTP, database, and queue tabs appear when the corresponding OpenTelemetry semantic attributes are present. See OpenTelemetry integration to start sending traces.