When a backup is overdue or failed

You run these on your server. Ballast doesn't run them and can't: it only reads.

Overdue

The newest backup is older than its limit. Either the schedule stopped, or it runs less often than the limit expects. If it simply runs less often, change the limit on the backup's page.

Then check the schedule:

systemctl list-timers --all
sudo grep -r restic /etc/cron.d /etc/crontab
sudo crontab -l

Failed

Ballast read the backup, and its own record says the last run had a problem: a crashed run, errors in a snapshot, or a failing line in the log. Look at the log of whatever runs it:

journalctl -u your-backup.service --since "2 days ago"
grep CRON /var/log/syslog | tail -n 20

A borg .checkpoint archive means a run was cut off before it finished. Kopia's error count usually means files it wasn't allowed to read.

Can't check

Ballast couldn't look, which doesn't prove the backup failed. It reminds you every few days, because a backup could have stopped behind this. Each page shows the exact command it tried, so run that on the server as the same account to see the whole message.

• Permission denied: see Let Ballast read your backup.
• Wrong password: the password file moved or changed.
• Busy: another backup holds the lock, which clears by itself. Through borgmatic only.

No backup has run yet

The repository is there, but holds no snapshots. Check that the schedule exists and has reached its first run.