Softwr

Cloud · head to head

Thanos vs Velero

Thanos logo

Thanos

Cloud

Highly available Prometheus with long-term object storage

From
Free
Rated
-
Velero logo

Velero

Backup

Open source backup and migration for Kubernetes cluster resources

From
Free
Rated
-

The short version

  • Each has a real cost: Thanos several components — sidecar, store, querier, compactor, ruler — each with its own configuration and failure modes; Velero there is no vendor support contract from the project itself, so a failed restore during an outage is resolved by your team reading source code and GitHub issues.
  • They diverge on capability: Thanos covers Global query, Velero covers Cluster resource backup.
  • Prices and features above were last checked on 1 September 2026.

Where they differ

Only the attributes on which Thanos and Velero actually diverge.

Attributes where Thanos and Velero differ
AttributeThanosVelero
PlatformsKubernetes, Linux, DockerLinux
CategoryCloudBackup

Identical on both: starting price (Free), pricing model (Open source, no licence fee), free tier (Yes), user rating (Not yet rated).

What each one covers

Drawn from each product's published feature list. An absence here means we hold no record of it - not that the product lacks it.

Only in Thanos

  • Global query
  • Object storage retention
  • Deduplication
  • Downsampling

Only in Velero

  • Cluster resource backup
  • Volume snapshots
  • File-level backup
  • Cross-cluster restore
  • Backup hooks
  • Scheduled backups
  • Object storage targets

What people use each for

The jobs each tool is most often brought in to do.

Thanos

  • Querying metrics across many clusters or regions from one placenot Velero
  • Retaining metrics for years without local disk growthnot Velero
  • Removing the gap that appears when a single Prometheus instance restartsnot Velero

Velero

  • A platform team needing namespace-level restore after an accidental deletion, with no budget for a backup productnot Thanos
  • Migrating workloads between Kubernetes clusters or between cloud providers using backup and restorenot Thanos
  • Capturing cluster configuration before a risky upgrade so the previous state can be reinstatednot Thanos
  • An organisation proving a Kubernetes recovery capability to an auditor before deciding whether to buy a commercial productnot Thanos

Where each one falls short

Documented limitations, not opinions. Every one is a constraint you would hit in normal use.

Thanos

  • Several components — sidecar, store, querier, compactor, ruler — each with its own configuration and failure modes
  • The compactor is a common source of operational trouble and must not run twice against the same bucket
  • Query latency over object storage is meaningfully higher than local Prometheus
  • Object storage costs and API request charges become real at high volume

Velero

  • There is no vendor support contract from the project itself, so a failed restore during an outage is resolved by your team reading source code and GitHub issues.
  • Application consistency depends on backup hooks you write and maintain, and a database backed up without a correct hook restores to a state that may not be usable.
  • Restore ordering across dependent resources such as custom resource definitions and operators needs manual attention, and complex applications frequently need a documented runbook rather than a single restore command.
  • Corporate engineering investment became uncertain after Broadcom acquired VMware, which is a real consideration for anyone building a long retention policy on the project.
  • Object storage egress charges apply on restore and are billed by your cloud provider, so a full cluster recovery from another region carries a cost discovered during the incident rather than during planning.

Pricing, plan by plan

Thanos

Free
  • ThanosFree
    • Full functionality
    • No usage limits
    • Community support

Velero

Free
  • VeleroFree
    • Apache 2.0 licensed
    • No cluster or capacity limits
    • Community support via Kubernetes Slack and GitHub

Which should you pick?

Choose Thanos if

  • You need global query.
  • You want to start without paying.
  • You work on Kubernetes, Linux, Docker.
  • You also want object storage retention.

Choose Velero if

  • You need cluster resource backup.
  • You want to start without paying.
  • You work on Linux.
  • You also want volume snapshots.

Questions people ask

Is Thanos or Velero better?
Neither clearly leads. Thanos starts at Free and Velero at Free, and user ratings are close enough to be indistinguishable. Choose on capability and platform support.
Which is cheaper, Thanos or Velero?
Thanos starts at Free and Velero at Free.
Does Thanos or Velero run on more platforms?
Thanos runs on Kubernetes, Linux, Docker. Velero runs on Linux.
Can I use Thanos for free?
Both have a free tier, so you can try either at no cost before committing.
What is Thanos best used for?
Thanos is most often used for querying metrics across many clusters or regions from one place, retaining metrics for years without local disk growth, removing the gap that appears when a single prometheus instance restarts. Of those, querying metrics across many clusters or regions from one place and retaining metrics for years without local disk growth are not what Velero is typically brought in for.
What can Thanos do that Velero cannot?
Thanos covers Global query, Object storage retention, Deduplication, Downsampling. Velero covers Cluster resource backup, Volume snapshots, File-level backup, Cross-cluster restore.

Answered from the vendors’ own pages

Thanos: Is Thanos free?

Yes, open source and CNCF-incubating. Costs are the object storage it uses.

Velero: Can we buy support for Velero?

Not from the project. Several commercial Kubernetes backup products package or extend it and sell support, which is the practical route to an accountable vendor.

Thanos: Does Thanos replace Prometheus?

No. It runs alongside existing Prometheus servers, adding global query, deduplication and long-term storage.

Velero: Does it back up persistent volume data?

Yes, through cloud or Container Storage Interface snapshots, or through file-level upload where snapshots are not available.

Thanos: Thanos or VictoriaMetrics?

Thanos layers onto Prometheus using object storage and is the more established multi-cluster answer. VictoriaMetrics is a separate store aiming at lower resource use and fewer moving parts.

Velero: What does it cost?

Nothing to licence. You pay for object storage capacity and for egress on restore, both billed by your cloud provider.

Share

Related pages

Other head to heads