Removing a Cluster

1. Uninstall the Helm chart

helm uninstall dataagent --namespace dataagent

This removes the agent pod, its service accounts (including the dataagent-live-read service account used for live lookups), the ClusterRoles and ClusterRoleBindings, and the bundled monitoring pods (prometheus-server, kube-state-metrics, and node-exporter).

The mTLS credential secret (dataagent-mtls-credentials) is created at runtime by the agent, not by Helm, so it survives helm uninstall. Delete it explicitly to avoid a reinstall reusing the stale certificate:

kubectl delete secret dataagent-mtls-credentials --namespace dataagent --ignore-not-found

2. Delete the namespace (optional)

The dataagent namespace is always preserved on helm uninstall, regardless of who created it. To remove it:

kubectl delete namespace dataagent

3. Remove the cluster from the dashboard

After uninstalling the agent, the cluster will appear as Disconnected in the dashboard. To remove it from the dashboard, archive it (requires the Admin or Owner role):

1. Go to Clusters
2. Find the disconnected cluster and open its ⋯ menu
3. Click Archive and confirm

Effect on incidents

• Active incidents from an uninstalled cluster are automatically resolved when the agent disconnects and stops sending heartbeats (within 4 hours via auto-resolution).
• Archiving the cluster hides its active incidents from the dashboard.