Skip to main content

Backup & Restore

Overview​

Modern applications are composed of multiple microservices running in containers, and Kubernetes has no built-in backup or recovery policies by default. Cloud Orbiter integrates Velero as a first-class Add-On to provide robust backup, restore, and disaster recovery for Kubernetes workloads.

Velero is a battle-tested open-source tool that takes backups of Kubernetes cluster resources and persistent volumes, restores them, and supports cluster migration.

What Cloud Orbiter Backup Provides​

  • Namespace Backup — Back up entire namespaces and restore them as required
  • Persistent Volume Backup — Snapshot and restore PVs alongside namespace resources
  • Cluster Resource Migration — Move workloads from one cluster to another
  • Environment Replication — Replicate production clusters into development or testing clusters
  • Disaster Recovery — Restore cluster state after failure or data loss

Architecture​

Cloud Orbiter uses Velero with an S3-compatible storage backend. Every cluster backup is stored in a configured S3 endpoint, providing durable, geo-redundant backup storage.

Kubernetes Cluster → Velero Agent → S3-Compatible Storage Endpoint
                                         ↕ (restore)
                    Target Cluster ← Velero Agent ←

Prerequisites​

  1. Velero Add-On enabled on the cluster
  2. S3 Storage Endpoint configured in Cloud Orbiter

S3 Storage Endpoint Onboarding​

Before creating backups, configure the S3 storage endpoint where backup data will be stored:

  1. Navigate to Backup & Restore → Storage Endpoints
  2. Click + Add Endpoint
  3. Provide:
    • Endpoint name
    • S3 URL (bucket URL or compatible storage URL)
    • Access Key and Secret Key
    • Region and bucket name
  4. Click Save — Cloud Orbiter validates the connection

Creating a Backup​

  1. Navigate to the cluster dashboard
  2. Click Backup & Restore → + Create Backup
  3. Configure the backup:
    • Backup Name — unique identifier for this backup
    • Namespaces — select specific namespaces or all namespaces
    • Storage Endpoint — select the S3 endpoint configured above
    • Schedule — one-time or scheduled (cron-based)
  4. Click Create to start the backup job

Restoring from Backup​

  1. Navigate to Backup & Restore → Backup Jobs
  2. Select the backup you want to restore from
  3. Click Restore
  4. Configure restore options:
    • Target cluster (same or different cluster)
    • Target namespace(s)
    • Resource inclusion/exclusion options
  5. Click Restore — Velero applies the backup to the target cluster

Backup Resource Customization​

Cloud Orbiter supports granular backup configuration to optimize storage usage and backup performance:

Namespace Selection​

Include or exclude specific namespaces from the backup. Use label selectors to target namespaces with specific attributes.

Persistent Volumes​

  • Include — Back up specific PVs using label selectors (e.g., label data=true)
  • Exclude — Skip PVs that are larger than a size threshold or managed by a separate backup process

Snapshot Volumes​

  • Include or exclude volume snapshots from backup operations based on your storage strategy

Cluster-Wide Resources​

Include or exclude cluster-level resources (e.g., nodes, ClusterRoles) when only application-specific resources are needed.

Label Selectors​

Target resources for backup or exclusion using Kubernetes labels:

  • Include all resources labeled tier=backend
  • Exclude resources labeled backup=false

App Hooks​

Add pre/post-backup scripts to ensure data consistency:

  • Pre-backup hook — Run a script to quiesce a database before backup starts (e.g., stop writes, flush cache)
  • Post-backup hook — Run a script to resume normal database operations after backup completes

Example use case: Stop a PostgreSQL service → take backup → restart the service — ensuring a consistent, crash-free backup.

Backup Job Status​

StatusMeaning
In ProgressBackup is currently running — resources being serialized and uploaded to S3
CompletedBackup finished successfully — available for restore
FailedBackup encountered an error — review logs for details
PartialSome resources backed up successfully; others failed — review logs

Cluster Migration​

Cloud Orbiter backup enables seamless cluster migration:

  1. Create a full backup of the source cluster
  2. Provision a new target cluster
  3. Restore the backup to the target cluster
  4. Validate application health on the target
  5. Switch traffic to the new cluster

This workflow supports migrations between:

  • On-premise to public cloud
  • Different cloud providers
  • Different Kubernetes distributions
  • Production to staging/dev environments