been reading "how complex systems fail" by richard cook, 18 short points from a doctor who studied why things break in hospitals. it was written in the late 90s and it reads like a postmortem of every outage i've ever been part of.
the one that hit me: there's no such thing as a root cause. catastrophe needs several small failures lining up, each harmless on its own. "root cause" is mostly our social need to point at someone. yesterday the us grounded every flight for hours and the official answer was "old equipment." sure. and also everything else that quietly had to go wrong at the same time.
my other favorite: complex systems always run in degraded mode. everything is a little broken all the time, and it works because people keep patching it on the fly. that's not a bug report, that's my homelab's biography.
the uncomfortable takeaway is that safety isn't a part you buy. it's something people create every day, and the "fix" after an accident usually just adds more complexity for the next one.

