Backups

Miabi backs up both your databases and your volume data, on demand or on a schedule, and ships the result to a backup target. Database backups are built on the proven ecosystem tools pg-bkup (PostgreSQL) and mysql-bkup (MySQL / MariaDB), each run as a one-shot container against your database.

What you can back up

Type	Engine / source	Tool
Database backup	PostgreSQL	pg-bkup
Database backup	MySQL / MariaDB	mysql-bkup
Database backup	MongoDB	mongodb-bkup
Database backup	libSQL	libsql-bkup
Volume archive	Any workspace volume	volume-bkup (S3 only)

Redis backups

Redis backups are not yet supported. For Redis, rely on persistence and replication for now.

See Databases overview for how databases are provisioned in a workspace.

Manual backups

A manual backup runs once, immediately.

1. Go to Storage → Backups (or open the database / volume and use its Backups tab).
2. Select Create backup.
3. Choose the source (a database or a volume) and the backup target that receives the result.
4. Run it. Progress and the final status appear in the backup history.

Scheduled backups

Scheduled backups run automatically on a cron expression, so you get hands-off, recurring protection.

1. Select Create schedule on the backup you want to automate.

2. Pick the source and the destination backup target.

3. Enter a cron expression for the cadence, for example:

   Cron	Runs
   0 2 * * *	Every day at 02:00
   0 */6 * * *	Every 6 hours
   0 3 * * 0	Weekly, Sunday 03:00

4. Save. Each run is recorded in the backup history with its outcome.

tip

Stagger schedules (for example databases at 02:00, volume archives at 03:00) so heavy jobs don't contend for CPU and I/O at the same moment.

Volume archives

Volume archives capture the full contents of a volume as a single archive and send it to the workspace's backup target.

Volume archives require S3

There is no local destination for volume archives. Configure an S3 (or S3-compatible, e.g. MinIO) backup target for the workspace first, or the archive fails with workspace S3 backup settings are not configured. Volume archives also run on demand only — scheduling is available for database backups, not for volumes.

Restoring

1. Open the backup history and find the backup or archive you want.
2. Select Restore.
3. Choose the target database or volume to restore into and confirm.

Restore runs as a one-shot job and reports its status when finished.

caution

Restoring overwrites existing data in the selected target. Confirm you have the right backup before proceeding, and prefer restoring into a fresh database or volume when validating a backup.