The **Labels** tab analyzes label-level impact on storage and cardinality.

- Lists all labels attached to the metric
- Shows each label's usage, cardinality, and unique value count
- Highlights labels driving high series counts

Note

Watch for labels like `pod_name` or `operation` with many unique values; they often inflate cost.

## Explore label details

From the **Labels** tab, click any label row (for example, `operation` or `host_name`) to open its detailed view.

This view helps you analyze how a specific label contributes to metric usage, variation, and value diversity.

The label details page includes two sub-tabs:

- **Metrics:** Lists all metrics associated with the selected label
- **Live values:** Displays recent values for that label

Use this view to understand which metrics and label values drive data volume and cardinality.

## Metrics tab

The **Metrics** tab shows how the selected label is used across different metrics.

For each metric, you can review:

- **Usage:** Data volume attributed to the label
- **% Usage:** Relative share of total usage
- **Unique Values:** Count of distinct label values
- **Last Ingested:** Timestamp of the latest ingestion for that metric

Example: Selecting the `operation` label lists metrics such as `cx_service_catalog_service_duration_cx_bucket` and `latency_ms_sum`, showing how each contributes to total usage.

Use this view to detect metrics that are heavily influenced by the chosen label or that have excessive label variations.

## Live values tab

The **Live values** tab lets you explore the actual values for a selected label.

- Displays a sample of up to 1,000 values at a time. You can narrow the results using the search bar to filter by specific strings or patterns.
- "Live" refers to a best-effort view of the last 24 hours of data. The list reflects values recently observed in your environment rather than a complete historical record.
- Values are aggregated across all metrics that include the selected label, not limited to a single metric. Use this view to identify active label values and verify that expected data is still being ingested.

Note

Only 1,000 values are shown. Not all values are displayed.

**Use this view to:**

- Inspect label values to understand what data is associated with a label.
- Spot noise or garbage values in high-cardinality labels that may indicate misconfigured instrumentation.
- Confirm label consistency, such as checking that values follow expected formats or naming patterns (for example, `region: us-east-1`, not `region: useast`).

**Example:** If you select the label `operation`, you might see hundreds of API operation names such as: `GET /items`, `POST /checkout`, `grpc.health.v1.HealthCheck`, etc.