Yeah, us-east-1 is pretty good at failing already. We lost us-east-1c for most of the day about a week ago due to a fiber line being cut. I'd estimate that AWS

Yeah, us-east-1 is pretty good at failing already. We lost us-east-1c for most of the day about a week ago due to a fiber line being cut. I'd estimate that AWS manages fewer than "three 9s" in us-east-1 on average. Not across the board, but at any given time something has a decent chance of not working, be it an entire AZ, or regional S3, etc. They're still pretty reliable, and I like the idea of a zone with built-in failure for testing things, but your joke about us-east-1 is based in solid fact.