Kubernetes Disaster Recovery & Backup with Velero

Implement enterprise disaster recovery, scheduled backups, and cross-cluster migration using Velero. Learn how to back up cluster metadata and persistent volume snapshots to cloud object storage (AWS S3, MinIO, GCS), schedule automated retention policies, and recover entire namespaces after catastrophic failures.

flowchart TD
    subgraph LiveCluster["Primary Kubernetes Cluster"]
        Workloads["Pods, Services, Deployments, Secrets"]
        PVC[("PersistentVolumeClaims (CSI Disks)")]
        VeleroPod["Velero Server Pod (Namespace: velero)"]
    end

    subgraph ObjectStorage["Remote Object Storage (AWS S3 / MinIO)"]
        S3Bucket[("S3 Bucket: kubedaily-cluster-backups<br/>(Gzipped JSON metadata)")]
        Snapshots[("Cloud Storage Volume Snapshots")]
    end

    subgraph RecoveryCluster["Disaster Recovery / Target Cluster"]
        NewVelero["Velero Restore Process"]
        RestoredWorkloads["Restored Pods & StatefulSets"]
        RestoredPVC[("Restored Persistent Volumes")]
    end

    VeleroPod -->|1. Stream metadata tarball| S3Bucket
    VeleroPod -->|2. Trigger CSI Snapshot| Snapshots

    S3Bucket -->|3. Restore metadata| NewVelero
    Snapshots -->|4. Restore disks| RestoredPVC
    NewVelero --> RestoredWorkloads
    RestoredPVC --> RestoredWorkloads

1. Velero Architecture

Velero operates inside your cluster via a single controller deployment:

• Backup: Serializes Kubernetes API objects into compressed JSON tarballs and writes them to remote object storage while triggering storage volume snapshots via CSI.
• Restore: Reads backup archives from object storage and creates the resources in the destination cluster, matching PersistentVolumes to their point-in-time snapshots.

2. On-Demand and Scheduled Backups

Step 2.1: Create an On-Demand Backup of a Namespace

# Back up the entire 'production' namespace including persistent volumes
velero backup create prod-backup-$(date +%F) \
  --include-namespaces production \
  --snapshot-volumes \
  --wait

Step 2.2: Check Backup Progress and Summary

velero backup describe prod-backup-2026-11-01 --details

Output highlights:

Phase:  Completed
Namespaces:
  Included: production
Resource List:
  apps/v1/Deployment: 6
  v1/PersistentVolumeClaim: 4
  v1/Secret: 12
  v1/Service: 6
Persistent Volumes:
  pvc-9a8234ab: Snapshot created successfully

Step 2.3: Author Automated Scheduled Backups

Create a Schedule resource to run automated nightly backups with a 30-day retention period:

apiVersion: velero.io/v1
kind: Schedule
metadata:
  name: nightly-prod-backup
  namespace: velero
spec:
  # Runs at 02:00 AM every night
  schedule: "0 2 * * *"
  template:
    includeNamespaces:
      - production
    snapshotVolumes: true
    # Automatically delete backup after 720 hours (30 days)
    ttl: 720h0m0s

Apply:

kubectl apply -f nightly-backup-schedule.yaml
velero schedule get

3. Simulating a Catastrophic Failure & Restoring

Disaster Simulation

Accidentally delete the production namespace:

kubectl delete namespace production
# All deployments, services, secrets, and data are gone!

Emergency Restore

# 1. List available backups stored in S3
velero backup get

# 2. Trigger restore from the latest backup
velero restore create --from-backup prod-backup-2026-11-01 --wait

# 3. Check restore status
velero restore describe prod-restore-20261101...

Output:

Phase: Completed
Warnings: 0
Errors:   0

Verify with kubectl get all -n production: All workloads, secrets, and persistent volumes are completely restored and operational!

Summary Checklist

• Configure remote S3-compatible object storage for off-site backup storage.
• Enable volume snapshots via CSI plugins to guarantee application data recovery.
• Schedule automated nightly backups with explicit retention (ttl: 720h).
• Conduct regular disaster recovery drills to validate restore procedures and recovery time objectives (RTO).