For the complete documentation index, see llms.txt.
Skip to main content
Version: 8.10 (unreleased)

Restore troubleshooting

Camunda Enterprise

Use this guide to diagnose and recover from restore failures.

Common restore phases

Restore operations pass through these phases:

  • VALIDATING
  • DeleteData
  • EnterRestoreMode
  • RestoreSnapshots
  • ExitRestoreMode

Failure modes and actions

PhaseTypical symptomLikely causeRecommended action
VALIDATINGRestore request fails quickly with validation errorBackup is not complete, backup is legacy (missing generationUuid), partition count mismatch, or restore already in progressSelect a completed non-legacy backup, confirm partition compatibility, and retry after any active restore reaches terminal state.
DeleteDataRestore fails after initiation and cluster returns to previous stateInternal destructive phase could not complete safelyRetry once after verifying cluster health; if repeated, contact support with restore ID and timestamps.
EnterRestoreModeCluster remains unavailable and restore does not progressController/operator state transition did not completeCheck cluster status history and activity context, then contact support with restore ID.
RestoreSnapshotsRestore progresses but fails before completionSnapshot restore failure, backend storage issue, or compatibility issueRetry with a different backup if available. If all fail, contact support and provide failed restore IDs.
ExitRestoreModeRestore appears complete but final healthy state is delayed or failsPost-restore reconciliation did not completeWait for reconciliation window, then re-check cluster health. If still degraded, contact support with restore metadata.

API error reference

CodeMeaning
400Invalid restore input (for example, legacy backup, backup not complete, compatibility check failed)
404Cluster or backup was not found
409Another restore is already in progress
501Restore feature flag is not enabled
500Internal server or Kubernetes API error

Activity and audit interpretation

For each restore, capture:

  • Who initiated the restore
  • Which backup was selected
  • Start and completion timestamps
  • Final terminal state (COMPLETED, FAILED, or ABORTED)

Use this record for incident timelines and post-incident review.

Common mistakes

  • Restoring a backup without confirming partition count compatibility.
  • Starting restore during other cluster mutation activities.
  • Assuming restore is non-destructive for current cluster data.
  • Running restore without a validated rollback/runbook plan.

When to contact support

Contact support when:

  • Restore repeatedly fails with the same backup and a newer valid backup is not available.
  • Cluster does not return to healthy status after terminal restore state.
  • You suspect data integrity issues after a reported successful restore.

Include restore ID, cluster ID, selected backup ID, timeframe, and screenshots of status and errors.