Docs · Operations

Backups & recovery

Backups run without you scheduling anything. You manage them in your project under Backups and Settings.

Nightly and on-demand backups

• A backup of your database runs every night.
• A copy of each nightly backup is kept off the server for 30 days. To use one, fetch it back to the server, then restore it from the backups list.
• You can also create a backup on demand (for example, before a big migration) and restore it from the same list.

Point-in-time recovery

Point-in-time recovery lets you rewind the database to a chosen moment, for example just before a bad deploy or an accidental delete.

• Changes are archived continuously, at most about five minutes behind.
• You can restore to any second within the last 7 days.
• A restore keeps the data it replaces on the server, so you can still look at it afterwards.

Standby replica

On plans that include it, you can add a standby read replica: a second server that streams a continuous copy of your database.

• Apps on the same private network can use it as a read-only Postgres.
• If you need to, it can be promoted to a writable database.

Good habits

• Create an on-demand backup before large schema changes or imports.
• Ask McManus to explain a change before you apply it. Every change it proposes waits for your approval.
• Keep your schema in version control so you can rebuild it anywhere.