Deploy a Cron Job to Backup PostgreSQL to Google Cloud Storage

This repo can be used to deploy a Cron Job on Render to backup a PostgreSQL instance to Google Cloud Storage.

Fork this repo and click the button below to deploy.

See the guide for S3 at https://render.com/docs/backup-postgresql-to-s3 for more information.

Excluding tables from the backup

Set the optional EXCLUDE_TABLE_DATA environment variable to a comma-separated list of tables whose data to skip. The schema is still dumped — only the rows are dropped — so you can shrink backups by leaving out large, non-critical tables (caches, logs, etc.). Leave it unset to back up everything.

Values are passed to pg_dump --exclude-table-data, so wildcards and schema qualification work:

EXCLUDE_TABLE_DATA=temp_*, public.audit_log, event_stream

An unqualified name matches that table in every schema; qualify it (public.audit_log) to pin one schema.

Known limitations

• Typos pass silently. A pattern that matches no table is ignored, so a misspelled name just backs up that table's data as usual. Check names against the live schema.
• Can break restores via foreign keys. If a kept table references an excluded one, its rows lose their targets and the foreign key may fail to restore. Only exclude tables nothing else references.
• Doesn't cover partitions. Excluding a partitioned table doesn't exclude its child partitions. List the children explicitly (or use a wildcard).