Create a fully managed, highly available database from the Deploy Managed Database page in my.cubepath.com. CubePath runs the infrastructure, replication and failover for you.
NoteYour account must be verified before you can deploy.
Choose an engine
| Engine | What it is | Replication |
|---|---|---|
| PostgreSQL | Advanced open-source relational database | Streaming replication with automatic failover |
| MySQL | The world's most popular relational database | Group Replication (MGR), needs a quorum of 3+ nodes |
| Valkey | Redis-compatible in-memory store for caching, queues and real-time workloads | Primary / replica |
ImportantVersions are pinned. There are no in-place major upgrades, so the version you pick is the version that cluster runs. Moving to a new major means deploying a new cluster and migrating.
Deploy
- 1Pick engine and versionChoose PostgreSQL, MySQL or Valkey, then the version.
- 2Name your databaseGive it a name (lowercase letters, numbers and dashes), unique within your organization.
- 3Choose region and planPick a region, then a plan. Prices are per node, so the total scales with the replica count.
- 4Set replicasSelect the cluster size. PostgreSQL and Valkey start at 1 node; replication needs at least 2. MySQL requires at least 3.
- 5Configure and deployOptionally set a backup schedule (5-field cron) and retention under Advanced Settings, then click Deploy Database.
ImportantProvisioning a database cluster can take 30–45 minutes. The connection endpoint and credentials become available once it finishes.
NoteBackups are coming soon. The schedule and retention you set are saved with the database and take effect automatically once backups launch. Until then, take your own dumps — a managed cluster protects you from hardware failure, not from a bad migration.
Connect
Open the database and go to the Credentials tab, then click Reveal credentials. You'll get the host, port, username, password and a ready-to-use connection URI. Point your application at the URI or the host/port pair — there is no SSH access to the nodes.
WarningAnyone with these credentials has full access to your data. Create per-application users rather than sharing the admin account, and use Rotate credentials if the admin password is ever exposed.
Scale
Once the cluster settles into active (or degraded), Scale Database offers two independent moves — one at a time:
- Replicas (horizontal). Add or remove nodes between the plan's min and max. Nodes are added or removed without downtime, though the operation takes several minutes. Each replica is a separate billable node.
- Plan (vertical). Move the whole cluster to a bigger or smaller plan. Every node is resized with a rolling restart, so connections reconnect briefly.
The engine and version never change during a scale operation.
NoteA plan can't support fewer replicas than you're running. To move to a smaller plan, scale the replica count down first.
Statuses
Provisioning → Active is the normal path. You'll also see Updating (a configuration or credential change), Scaling, Degraded (running, but not at full redundancy — a node is down or rejoining) and Error.
Most operations are blocked while another one is running; the dashboard tells you to try again once it settles.
Protection and deletion
Enable Protection blocks accidental deletion. Deleting permanently destroys the database and all its data, so the dialog asks you to type the database name to confirm. Export anything you need first — there is no undo, and (until backups launch) no copy to fall back on.
The database tabs
Configuration, Credentials, Databases, Users, Metrics and Activity.