The error "Main slot of revision (number) not found in database!", documented in T212428, has been present for some time.
While this error should be fixed, the current logging is problematic for train triage, conductors, and deployers, who have to change how it is filtered on a weekly basis. Revise the logging to be more straightforward to filter.
From a comment on the original task: "How about a log call to a custom channel for the revision ID and a generic exception message for error logs? The latter would be much easier for deployers to filter / ignore."