Set a backup retention policy
Without a retention policy, backups either pile up forever (expensive) or you keep too few and can’t recover from a problem you discover late. The answer is a tiered policy: lots of recent snapshots, fewer older ones.
A sensible default
Keep:
- 7 daily — recover from anything in the last week.
- 4 weekly — a month of weekly checkpoints.
- 6–12 monthly — longer history for slow-to-surface problems.
restic’s retention (its forget + prune step) keeps the snapshots that match
your policy and deletes the rest, then reclaims the freed space.
Why tiered?
Some problems are found immediately (bad deploy → restore yesterday). Others take weeks to notice (a slow data corruption, a breach discovered late). Daily snapshots cover the first; monthly ones cover the second — without storing every single day forever.
Match it to your data
- Busy site / changing data → more frequent backups, more dailies.
- Rarely-changing site → fewer snapshots are fine.
- Compliance needs? Some businesses must retain backups for a set period — extend the monthly/yearly tier to match.
Cost vs. safety
restic deduplicates, so extra snapshots of mostly-unchanged data cost little. Err toward keeping more history than you think you need — storage is cheap compared to not having the snapshot you need.
Set it up
Configure retention alongside your offsite backups. Then make sure you can actually restore, and fold it into your disaster-recovery plan.
Running OrbixPanel? restic retention prunes old snapshots automatically. OrbixPanel.com → — the free, self-hosted, AI-powered hosting control panel.