Currently all transcoding jobs are logged to the transcode table. In case of an error, the error is logged to transcode_error, but there's no indication on which video scaler the error occurred.
That would be especially helpful for staged upgrades, e.g. when deploying a new ffmpeg release (e.g. for sec updates) where one canary server is upgraded ahead of the rest, as it would allow to monitor errors from that particular server.