Ask HN: What’s your “don’t forget to check the power cord” advice?

I once read about old school sysadmins keeping a large dummy file that was a few percent of the total disk size (e.g. 2GB file on a 100GB disk). If a server ran out of disk in the middle of the night due to a bad config, logs filling up, etc., one could delete the file and create some troubleshooting time.