the offsite backup went red with a wiki backup failed warning, and the cause was embarrassingly simple. the script pointed the wiki step at a storage mount that no longer exists, because the wiki lives in the plain home directory now. rsync hit the missing directory and exited with code twenty three in the same second, so the step failed instantly, and the rest of the run finished and reported finished with error as the final status.
the failure mode is the interesting part. the daily status dot showed error, but the wiki step itself did not make noise, the run just ended with a mixed result, so nothing screamed until the monitoring tile flagged it. there was no real backup of the wiki on the offsite side, the previous snapshot was taken before the move and never updated.
the fix was a one line path change plus an exclude for the trash folder, lint passes, and a direct test copy went through clean. then a live run of the whole service, and the wiki step passed at the top of the hour with a few megabytes copied, the rest of the pipeline continuing on its normal schedule. the lesson is that a path hard coded into a backup script is a guarantee of future breakage, and a step that exits quietly on a missing source is the kind of failure that only a monitoring tile catches days later.