Picture this: your production Kubernetes cluster implodes—etcd fried, nodes offline, stateful workloads hemorrhaging revenue by the minute. You’ve got Velero backups stashed in S3, ticking compliance boxes. But restoring? That’s where the nightmare begins, even with Velero now under CNCF wings.
Velero’s shift to CNCF Sandbox, announced at KubeCon EU, hits Kubernetes backup teams right in the gut. It’s not some feel-good open-source win. For ops folks wrestling stateful apps, this screams architectural reckoning: governance versus true independence.
And here’s the kicker—Broadcom’s handing over the reins isn’t charity. It’s damage control after VMware’s acquisition left scars.
Why Real Teams Should Care About Velero’s CNCF Drama
Look, if you’re running databases or queues on K8s, Velero’s your go-to for declarative backups. CRDs for Backups, Restores, Schedules—they dance at the API layer, capturing namespaces, PVCs, RBAC like a portable cluster blueprint.
But. Vendor-neutral governance? Sure, CNCF means no more Broadcom solo roadmap hijacks. Maintainers from Red Hat, Microsoft join the party. Broadcom’s own words nail it:
“We really don’t want people to mistrust the open source project and believe that it’s somehow a VMware thing even though it hasn’t been a VMware thing for quite some time.”
Trust repair, check. Yet ops reality bites harder.
Vendor-independent ops? Nah. That bucket in the cloud still needs your IAM creds. No cluster, no etcd—no Velero magic.
Does CNCF Governance Make Velero Backups Bulletproof?
Short answer: no.
Velero reconstructs state through Kubernetes APIs. Brilliant for migrations—VKS to EKS, no sweat. But limits glare under pressure.
Take full cluster death. Backup data sits pretty in object storage. Restore? Needs a live target cluster, valid IAM to fetch manifests, functional control plane.
That table from the announcement? It spells doom:
| Axis | What Velero Controls | What Velero Depends On |
|---|---|---|
| Backup Definitions | CRDs inside cluster | etcd — gone if cluster is gone |
| Restore Logic | Velero controller + API server | Working target cluster |
| Metadata | Object metadata, resource specs | External object storage bucket |
| APIs | Kubernetes API layer ops | Cloud IAM for bucket access |
Network partition? IAM outage? External DBs untouched? You’re rebuilding manually.
Air-gapped MinIO setups? Compliance theater—restore still craves live endpoints.
The Hidden Control Plane Power Play
Velero started at Heptio—Beda and McLuckie’s shop, gobbled by VMware in ‘19. Broadcom era amps the paranoia: pricing shocks, license wars, VCF lock-in fears.
This CNCF drop? Counterpunch. Open Velero signals: our stack’s not all proprietary venom. Clean split—community Velero for backups, VKS stays Broadcom’s.
My unique take: it’s Docker redux. Remember Docker Inc. donating containerd to CNCF in 2017? Escaped monopoly gripes, handed runtime to community. Broadcom’s scripting the same escape for Velero—neutralize backlash, keep enterprise sales humming.
Bold prediction: expect Tanzu bits next. Broadcom’s playing long game, open-sourcing edges to protect the core.
But skepticism reigns. Governance fixes code drift. Ops? You’re still chained to clouds, creds, clusters.
Teams miss this daily. “Backup succeeded” lulls into false security. Reality: restore-time gaps devour hours—DNS, certs, external services untouched.
So, what’s the fix? Layer in bootstrappers like cluster API. Dual-region storage. Credential vaults that outlive outages.
Velero’s stronger community-side. But don’t ditch your drills.
Why Does Velero’s CNCF Move Echo VMware’s Past Sins?
VMware’s Heptio buy birthed Velero under corporate shadow. Broadcom amps it—perpetual licenses axed, bundles forced.
Market whispers: everything locks down.
CNCF entry flips script. Broader contributors steer. No unilateral breaks.
Yet operations lag. That S3 bucket? Still your cloud’s puppet strings. IAM? Provisioned outside Velero’s reach.
Historical parallel: like Percona forking MySQL tools post-Oracle. Community fork preserves, but deps persist.
Broadcom critiques its PR spin—“not a VMware thing” dodges acquisition baggage. It’s very much Broadcom now.
🧬 Related Insights
- Read more: Covenant: The Habit App That Burns Your Stake If You Bail
- Read more: Power BI: Dashboard Savior or Excel’s Gilded Cage?
Frequently Asked Questions
What is Velero and why CNCF now?
Velero’s Kubernetes-native backup tool using CRDs for state snapshots. CNCF Sandbox means community governance, not Broadcom solo control.
Does Velero CNCF make Kubernetes backups vendor-independent?
Governance yes, operations no—still needs cloud storage, IAM, live clusters for restores.
Will Velero replace tools like Kasten or Portworx?
Maybe for API-layer portability, but lacks storage-layer depth; best as complement.
Can Velero recover a dead Kubernetes cluster from scratch?
No—requires operational target cluster first.
