I’m at one of the latter, so I feel this in my bones. I’ve watched what should have been an innocent config change snowball into a pair of VM clusters shitting back and forth for 2 hours. Implemented strict change control that day. Kind of a pain, but the team learned a lot that day!

A typo in software development or other shell based work could completely ass womp a system in ways that could lose a company lots of money.

Oopsies on prod systems, even with an outage window, can really fuck shit up. Seemingly small mistakes can quickly snowball into systemwide outages.