Field notes3 sections · 2 min readBy Dezső Mező · Published October 3, 2026.mdOpsBackupPostgresRecovery
The setup below is the boring version that has survived real incidents — automated copies, off-site storage, and a calendar entry that actually runs the restore.
01Automate the copy, delete the ritual
Nightly dumps and weekly snapshots that run on a timer survive; procedures that depend on a person remembering do not. If the backup needs a human, it will eventually not happen on the night it matters.
02One copy must live somewhere else
A backup on the same host or the same provider account dies in the same incident. One off-site copy — different provider, different credentials — is the difference between an outage and a company event.
03Schedule the restore, not just the backup
A quarterly calendar entry: pick a copy, restore it to a scratch environment, open the app, check one real record. The first time this takes an afternoon; by the third time it takes twenty minutes.
What to take away
- Automate backups; manual rituals fail silently.
- Keep one copy off-site under different credentials.
- Test-restore quarterly; time it and write it down.
- Encrypt the dump — a backup is the whole database.
More from the lab
Browse all entries
Monitoring that pages someone, not the dashboard
What custom software actually costs
Custom software or off the shelf — the honest test
The first ninety days of a retainer, as it actually goes
Want this looked at on your own system?Start a conversation
