kubectl, through GitOps tools like Argo CD or Flux, or by writing your own controllers against the Kubernetes API. For the underlying resource model and field reference, see Configure database.
Request limited availability access
CWDB is enabled per CKS cluster through a feature flag managed by the CWDB team. To request enablement:- Contact CoreWeave through your customer Slack channel with your CKS cluster name and the namespace or namespaces you plan to use.
- Wait for confirmation that the CWDB add-on is active on your CKS cluster.
cwdbclusters.data.coreweave.com and cwdbbackups.data.coreweave.com listed in the output.
Resources
CWDB exposes two Kubernetes custom resources:CWDBCluster
A managed PostgreSQL database. Defines the topology (primary plus replicas), compute, storage, and access settings for one database. Restores from a backup are also configured on this resource, through
spec.postgres.restore.CWDBBackup
An on-demand backup of a database. Use it for known recovery points before risky operations.
Get started
About
Architecture, topology, observability, and the shared responsibility model.
Create a database
Verify prerequisites and deploy your first database from the Cloud Console or with
kubectl.Configure database
Tune topology, compute, storage, database name, and extensions. Includes the full field reference for all CWDB resources.
Manage database
Edit a database, resize storage, or delete a database after creation.
Connect to a database
Extract connection strings and route traffic through the pooler Services.
Manage backups
How automatic backups work, how to take on-demand backups, and how to restore from a backup into a new database.
Troubleshoot
Common issues during limited availability and how to escalate.