Skip to content

Failure Clusters

Failure Clusters automatically group similar test failures together. Instead of manually reviewing every failed case run, you can identify recurring patterns and fix root causes once — addressing multiple failures at the source.


Accessing Clusters

Open Clusters from the project sidebar. If you see a feature lock, your plan does not include AI failure clustering. Upgrade to Core or above to enable it.


Understanding the Cluster List

The cluster list shows all AI-grouped failure patterns, with:

  • Label — AI-generated name for the failure pattern
  • Summary — brief description of the cluster
  • Severity — critical, high, medium, or low
  • Status — pending (still being analyzed) or analysed
  • Failures — number of test failures in this cluster
  • First Seen / Last Seen — when the pattern was first and most recently detected

Use the Severity, Status, and Search filters to narrow down the list.


Viewing Cluster Details

Click a cluster to open its detail page. Two tabs are available:

Overview Tab

Shows the AI-generated root cause analysis and fix suggestions once the cluster has been analyzed. If the cluster is still pending, you'll see "This cluster is still being analysed."

Sample Failures Tab

Shows a sample of the failures grouped into this cluster, each with:

  • Case name
  • Suite name
  • Error message (in monospace)
  • Similarity percentage — how closely this failure matches the cluster pattern

Managing Clusters

Renaming a Cluster

Click the ... menu on a cluster row or the cluster detail page, then select Rename. Enter a new label in the dialog.

Merging Clusters

If two clusters describe the same root issue, merge them:

  1. Select both clusters using the checkboxes in the cluster list.
  2. Click Merge in the bulk actions bar.
  3. Choose which cluster to keep as the target.
  4. Confirm the merge.

INFO

Merging requires selecting at least 2 clusters.

Recalculating Centroid

If a cluster's membership looks incorrect after new failures are added, select Recalculate Centroid from the cluster menu. This re-runs the similarity calculation.

Deleting Clusters

Select one or more clusters and click Delete. Deleted clusters cannot be recovered.


Exporting Clusters

Click Export CSV on the cluster list page to download all clusters as a spreadsheet.


See Also