Your databases, in the project. Backed up, on a private network.
PostgreSQL, MySQL, MongoDB or Redis, ready in seconds on persistent storage, reachable by your services without going through the Internet.
At a glance
- Engines
- PostgreSQL, MySQL, MongoDB, Redis
- Connection
- Project private network
- Backups
- Daily or weekly, 1 to 60 copies
- Files
- Persistent volumes
A database, from creation to restore.
01
Five engines.
Pick the engine and name the database: the password is generated and the database is ready in seconds.
It sits on a persistent volume: data survives restarts and image updates.
02
The project's private network.
Your services reach the database by name, on the project's internal network. That is the default connection.
Access from outside (your machine, a GUI tool) is turned on explicitly, database by database.
03
The built-in explorer.
The Database tab lists tables and their structure, with a query editor: Ctrl+Enter runs it.
Enough to check a value or fix a row without installing a client.
04
Scheduled backups.
A dump every day or every week, at the hour you choose (UTC). You keep 1 to 60 copies: beyond that, the oldest goes.
Backups live on a volume dedicated to the project, separate from the database's own. Dumps are restored from the list.
05
Files that stay.
A service's file system starts from scratch on every redeploy. A persistent volume keeps a folder: uploads, SQLite, a CMS's content.
Everything else stays ephemeral: only mounted paths persist.
What it does not do yet.
Scheduling produces dumps of managed databases. Volume snapshots are started by hand.
No point-in-time restore: between two backups, data written is not protected by a copy. The maximum possible loss is the gap between two copies, 24 hours for daily, 7 days for weekly.
A volume is attached to one instance. With several replicas, keep shared data in the database or object storage.
A volume's size is chosen at creation. Leave some headroom.
Deleting a service also deletes its backups: export what you want to keep first.
Questions about data
The rest of the platform
A database, next to your code.
Create a project, add PostgreSQL: your services connect over the private network.