If a cluster disappears, recreating its Deployments will not necessarily bring back its database, uploaded files or credentials. A Kubernetes backup plan needs both the instructions for running applications and the data those applications depend on. Start by deciding what must be recovered and how long the application can be unavailable.

What a Kubernetes backup tool does

Velero backs up and restores Kubernetes resources and supports persistent-volume backup mechanisms. Database-native tools handle the database's own backup and recovery requirements. A storage snapshot captures a volume at a point in time, but whether that copy is usable by the application depends on how the application writes and how the backup was taken.

Use a cluster backup tool when you need to recover Kubernetes resources and supported volume data together. Keep database-specific backup procedures when you need database-aware recovery, and keep required credentials and encryption keys recoverable too. These approaches often complement one another. Test by restoring into a separate environment and using the application, not just by checking that the backup job completed.

The original record

Historical Kubedex content

Preserved for context. Commands, versions, prices and results below reflect the original research.

This page restores a surviving research workbook. It is a new location for historical source material, not a claim that this URL existed on the original site.

Historical workbook values; blanks mean unknown, not No. Prices, versions, maturity labels and feature claims are not current recommendations.

Kubernetes Cluster Backups / Backup (historical)

Historical snapshot

Recovered comparison data. Versions, prices and availability describe the original research, not a current benchmark. Blank or damaged source values are marked unknown.

Choose columns
Visible columns
Kubernetes Cluster Backups / Backup (historical)
ReferenceEtcd OperatorHeptio ArkKube-Backupetcd ctlkaptaindReShifter
Etcd deploymentEtcd cluster is deployed in a seed k8s cluster. Etcd could run on a k8s cluster or as a system daemon.UnknownUnknownUnknownUnknown
Etcd lifecycle managementEtcd operator creates the cluster and manages its lifecycle (resize, upgrade, failure recovery)Ark does not support etcd cluster deployment and lifecycleUnknownUnknownUnknownUnknown
Backup / restoreYesQuery apiserver to obtain all configurations and create a backup from that.UnknownUnknownUnknownUnknown
Backup methodEtcd snapshotUnknownUnknownUnknownUnknownUnknown
Automated Minority Failure Recovery (VM failure)Etcd-operator handles it automatically. It deletes failed pod and creates a new pod.No, if etcd cluster is running as k8s Deployment then it depends on Kubernetes to recover from pod failureUnknownUnknownUnknownUnknown
Automated Minority Failure Recovery (data corruption)Yes. It deletes old pod and create a new pod. Old data/PV is not used.NoUnknownUnknownUnknownUnknown
Majority Failure (VM failure)Yes. If PV is used for data, then new etcd pod will be created with the PV, so recovery will be automatically. If PV is not used for data then it requires manual intervention to recover from backup.Yes, requires manual intervention to recover from backupUnknownUnknownUnknownUnknown
Majority Failure (data corruption)Yes, requires manual intervention to recover from backupYes, requires manual intervention to recover from backupUnknownUnknownUnknownUnknown
Health MonitoringUses a liveness prob to check health of etcd pod.NoUnknownUnknownUnknownUnknown
Restore MethodManual using etcdctlManual using ark CLIUnknownUnknownUnknownUnknown
PV / Statefulset backupNoYesUnknownUnknownUnknownUnknown
Granular backup / restoreNoYesUnknownUnknownUnknownUnknown
Backup locationObject Store, PVObject Store, BlockUnknownUnknownUnknownUnknown
Garbage collectionYesYesUnknownUnknownUnknownUnknown
etcd cluster resizeYes, etcd-operator uses Custom Resource with replica count which can be modified to resize.NoUnknownUnknownUnknownUnknown
etcd cluster upgradeYesNoUnknownUnknownUnknownUnknown

16 rows

Sources & further reading

  1. Kubernetes Cluster Backups: historical workbook
  2. Velero backup and restore overview
  3. PostgreSQL backup and restore

Spotted something that needs another look?

Help improve this page →