Velero 0.11: Kubernetes Disaster Recovery Takes Flight

Introduction

On June 29, 2018, Heptio announced Velero 0.11, the new name and first release of the project formerly known as Ark. Velero packages backup, restore, and migration tooling for Kubernetes clusters, combining CRDs with pluggable object storage.

Core Capabilities

Backup & Restore Workflows

• Schedules full or incremental backups of cluster resources and persistent volumes.
• Stores snapshots in cloud object storage such as Amazon S3, Google Cloud Storage, or Azure Blob.
• Restores workloads into the same or different clusters, enabling rapid disaster recovery.

Volume Snapshot Support

• Integrates with cloud provider snapshot APIs (EBS, GCE PD, Azure Managed Disks).
• Supports restic-based file system backups for environments without native snapshots.
• Ensures volume backups coordinate with resource manifests for consistent restores.

Extensible Architecture

• Introduces a plugin framework for custom volume snapshotters and object store backends.
• Uses CRDs (Backup, Restore, Schedule) that fit naturally into GitOps pipelines.
• Ships with a CLI for ad-hoc operations and verification.

Getting Started

1. Install Velero CLI and server components:

   velero install \\
     --provider aws \\
     --bucket velero-backups \\
     --secret-file ./credentials-velero \\
     --backup-location-config region=us-west-2
   

2. Create a sample backup:

   velero backup create demo --include-namespaces kube-system
   

3. Verify the object storage bucket for compressed backup archives.

4. Restore into a fresh cluster with velero restore create --from-backup demo.

Practical notes (so restores actually work)

• Back up more than workloads: include cluster add-ons you rely on (Ingress, DNS configs, certificate managers) or document a clear restore order.
• Snapshots aren’t the whole story: PV snapshots help, but many apps still need application-level coordination (quiescing, flushing) to avoid inconsistent restores.
• Test restores regularly: the first time you run velero restore should not be during an incident.

Ecosystem Fit

• Complements Kubernetes 1.11 cluster upgrades by safeguarding state beforehand.
• Pairs with ExternalDNS 0.1 to preserve ingress endpoints across restores.
• Works alongside Prometheus 2.3 to capture monitoring configuration and alerts.

Summary

Velero 0.1 solidified Kubernetes-native backup and restore operations, empowering teams to treat cluster state with the same rigor as application code.