`should_force_full()` threshold in backup.py may be suboptimal for low-volume mail servers

[snrkl]

should_force_full() threshold may be suboptimal for low-volume mail servers

Environment:

• Mail-in-a-Box v73
• Ubuntu 22.04
• S3 backup target

Description:

should_force_full() seems to trigger a full backup on weekends only when cumulative incremental size exceeds 50% of the last full backup. On low-volume mail servers where daily incrementals are small relative to the full backup size, this threshold may never be reached in practice. Without a time-based fallback, the incremental chain can grow indefinitely against an increasingly old full backup.

Steps to reproduce:

1. Configure MIAB with S3 backup target on a low-volume mail server
2. Allow backups to run for an extended period
3. Observe that no full backup is triggered automatically despite the incremental chain growing over time

Potential fix:

In management/backup.py, should_force_full() around line 166:

 		if weekend:
 			if inc_size > .5*bak["size"]:
 				return True
+			if first_full_date + datetime.timedelta(days=90) < datetime.datetime.now(dateutil.tz.tzlocal()):
+				return True
 		return False

It would be helpful if the specific threshold value could be user-configurable. Ideally though the UI, or potentially through a custom.yaml for power users.

It is noted that:

• a custom YAML config likely goes against the design principles of the project
• The use case for smaller volume mail systems also likely occurs more with implementations that do not have significant engineering support to deploy 'under the hood' customisations, so a userland backup setting for "force full backup every N" might be an appropriate solution.