How to disable BACKUP_NUM_TO_RETAIN while keeping scheduled base backups?

[73rry]

Hi!

We are currently using the scheduled basebackup mechanism with BACKUP_NUM_TO_RETAIN as described here: https://postgres-operator.readthedocs.io/en/latest/administrator/#wal-archiving-and-physical-basebackups.

We are running into some issues with the built-in backup cleanup logic and are evaluating an alternative approach. Based on the comment below, we are considering delegating retention entirely to S3 lifecycle rules:

Number of backups is set by BACKUP_NUM_TO_RETAIN. Spilo sets it to 5 by default, no? You can override it in the manifest in the env section of a cluster. You can also define lifecycle rules on S3 buckets. This is how we clean up old Postgres log files for example.

— @FxKu in #2942

Setting BACKUP_NUM_TO_RETAIN to a very large number works as a workaround, but it isn't ideal.

Setting a very large BACKUP_NUM_TO_RETAIN value technically works but feels like a workaround. Is there a proper way to completely disable Spilo's cleanup logic while still keeping the scheduled basebackup feature enabled?

Thanks!