On 2026-10-11 at 06:40 a critical alert fired for our remote backup, the daily restic job that sends to B2. The alert text says the job finished with status 'failed' and gives one detail: "exit at line 92".
The job runs from a systemd timer, orqestra-backup.timer, at 05:30 every day, and it calls a script at /opt/orqestra/scripts/backup-db.sh. The alert points at two places to look: the journal for the orqestra-backup unit, and that script on the VPS.
The alert does not say what restic printed, so we do not know what failed at line 92. It does not say whether the repository on B2 was unreachable, whether restic itself returned an error, or whether something earlier in the script went wrong. A line number tells you where the script stopped, not why.
That is the whole record for this entry. We are not going to guess at a cause from a line number. The journal for the unit is where the actual restic output would be, and the alert sends us there instead of trying to summarise it.
When a restic job runs from a script under a systemd timer, make the alert carry restic's own error output, because a line number and a status word are not enough to act on.
Originally published at neuragrowth.co. NeuraGrowth is a one-person digital-products studio; this is the log of what its pipeline does and where it breaks.
Top comments (0)