Skip to main content

Metrics Analyzer

Metrics Analyzer gives you a complete picture of your metrics estate — cardinality, cluster breakdown, and which dashboards or alerts consume each metric. Use it to identify high-cardinality offenders, unused metrics, and optimization opportunities.

Metrics Analyzer — By Prefix view

Getting Started​

Navigate to Metrics → Analyzer in the sidebar. The page opens in the By Prefix tab by default.

Page Controls​

ControlDescription
ClustersFilter metrics to one or more clusters.
Time rangeChoose the analysis window (e.g. last 24 h).
Total cardinalityHeadline number showing total active time-series count across the selected scope.

Views​

By Prefix​

The default view groups metrics by their name prefix (the part before the first _). This is useful for understanding which subsystems or exporters contribute the most cardinality.

ColumnDescription
PrefixThe shared name prefix (e.g. container_, node_).
Total CardinalitySum of active series for all metrics under this prefix, with a bar chart showing relative size and cluster breakdown on hover.
Metric CountNumber of distinct metric names under the prefix.

Prefix depth — A dropdown above the table lets you control how many underscore-separated segments form the prefix (1, 2, or 3). Increasing the depth creates finer-grained groups.

Click any row to see the individual metrics within that prefix.

By Metric​

Switch to the By Metric tab to see every metric individually.

ColumnDescription
Metric NameFull Prometheus metric name.
CardinalityActive time-series count with a bar chart and cluster breakdown tooltip. Percentage of total is shown.
DashboardsNumber of Grafana dashboards that reference this metric. Click the count to view the list.
AlertsNumber of alert rules that reference this metric. Click the count to view the list.
ActionsA block button (visible on hover) lets editors create a drop rule for the metric directly.

Additional controls in this view:

  • Search — Filter the table by metric name.
  • Show only unused — Toggle to surface metrics that appear in no dashboards and no alerts, making them candidates for dropping.

Metric Detail Panel​

Click any metric row to open the detail side-panel with three tabs:

Stats​

  • Cardinality over time — A time-series chart showing how the metric's cardinality has changed, broken down by cluster (top 5 shown, rest grouped as "Others").
  • Label cardinality table — Lists every label on the metric together with the number of unique values for that label. This helps pinpoint which label is driving high cardinality.

Dashboards​

Lists all Grafana dashboards that reference the selected metric, with direct links to each dashboard.

Alerts​

Lists all alert / monitor rules that use the metric, with links to each rule.

Workflow: Reducing Cardinality​

  1. Open the Analyzer and sort the By Prefix view by Total Cardinality descending to find the largest contributors.
  2. Drill into a prefix to see which individual metrics are responsible.
  3. Click a metric to open the detail panel. Check the Label cardinality table to identify the high- cardinality label (e.g. a pod label with thousands of values).
  4. Check Dashboards and Alerts — if the metric is unused (both counts are 0), it is a candidate for dropping.
  5. Create a drop rule — Hover over the metric row and click the block icon, or go to Metrics Drop Rules to create a rule manually.

Best Practices​

  • Review regularly — Run a monthly review of the Analyzer to catch cardinality growth before it becomes costly.
  • Use "Show only unused" to find metrics that can be safely dropped.
  • Combine with Drop Rules — After identifying waste in the Analyzer, create a drop rule to prevent the metric from being stored going forward.

Support​

If you have any questions or need assistance, please contact us via our help chat app available on the Support link in the sidebar, or by reaching out to support@oodle.ai.