Your Supabase project was paused. Here is what that means for your data.

What happened

Supabase pauses a Free plan project that has not had enough database activity over the previous week. In its words, a few user requests to the database each day is usually enough to keep it active. Paid projects are never paused.

While paused, the database accepts no connections and the API returns errors. The storage volume is kept as it was. Nothing is deleted at this point.

You can restore a paused project yourself for up to one year after it was paused, from the project's page in the Supabase dashboard. After that window the project can no longer be restored in place; Supabase says a backup file of the project can be downloaded from the dashboard instead.

What to do right now

• Open the project in the Supabase dashboard and click Restore project. It takes a few minutes to come back.
• As soon as it is up, take a copy that lives outside Supabase. From your machine: pg_dump -Fc -h aws-0-<region>.pooler.supabase.com -p 5432 -U postgres.<project-ref> postgres > supabase.dump. Use the session pooler on port 5432; the transaction pooler on 6543 cannot run pg_dump, and the direct db.<ref>.supabase.co host is IPv6-only unless you pay for the IPv4 add-on.
• If the project was paused more than a year ago, download the backup file Supabase offers on the dashboard and restore it into a fresh project or any other Postgres with pg_restore.

So it does not happen again

If the project matters, upgrade it; paid projects are not paused. If it is a side project you want to keep on the free plan, a scheduled backup does two things at once: the daily connection counts as database activity, and you hold a copy that restores anywhere if the project is paused, deleted or migrated.

Dumpling runs pg_dump against the session pooler on a schedule and streams the file into a bucket you own, then checks with pg_restore --list that it would actually restore. One database with daily backups is free. The setup is in the guide linked below and takes about fifteen minutes, most of it creating the bucket.

# restoring a Supabase dump anywhere
createdb restored_db
pg_restore --no-owner --no-acl -d restored_db supabase.dump