Profiles Services View

The Profiles landing page lists every service that sent profile samples in the selected time range, sorted by its share of total samples. Start here to find which services are worth investigating.

Columns

Column	Description
Service	Service name, from the service.name resource attribute. Click it to open the flamegraph.
Env	Deployment environment, from deployment.environment.name. Shown when the environment selector is All.
CPU share	The service’s share of all CPU samples across listed services in the selected range.
Runtime	Language runtime, from telemetry.sdk.language, when the profiler reports it.
Last profile	When the most recent sample for the service arrived.

CPU share is relative, not absolute. A service at 59% is using more than half of the CPU sampled across the profiled services, not 59% of a core.

Narrowing the List

Use the controls at the top of the page:

• Environment selector: limit results to one environment.
• Time range: choose the window to aggregate samples over.
• Search services: filter the list by service name. Press / to focus the search box.

Filters

The Filters panel on the left narrows results by profile attributes:

Filter	Source attribute	Notes
Profile Type	Profile sample type	CPU, Alloc, or Wall. Types with no data in the current range are disabled.
Cluster	k8s.cluster.name	Set by clusterName in the Helm chart values.
Namespace	k8s.namespace.name	Kubernetes namespace of the profiled pod.
Runtime	telemetry.sdk.language	Shows No runtime values when the profiler does not report a runtime.

---

Troubleshooting

• Filters look empty

  • Runtime shows No runtime values. The profiler does not report telemetry.sdk.language for these services. Other filters still work.
  • Alloc or Wall is disabled. No samples of that type arrived in the selected range.

• Services are missing or the list is empty. See Troubleshooting in the Profiles overview.

Please get in touch with us on Discord or Email if you have any questions.