This is an important point. My customers don't love outages (who does?) but I've had them and it doesn't really hurt that badly. My products aren't that critical. They're understanding as long as you communicate.
github took down our eng dept probably half a dozen times last year. I don't even see how they hit 3 nines last year. It was insane. Even before that we had to switch off Azure. That was a sad joke of a service. They must be running their entire ops out of a log cabin connected with the bare minimum of cabling and bandwidth necessary.
That's also the thing the serverless/API-all-the-things/cloud promoters don't get. This interconnected web of services is incredibly fragile. It's to the point that every day is a new failure. One day github is down. Next day your CI process breaks because Docker shits the bed. Next day your E2E service is hosed. Following day you hit an API limit and need to go dump more money into the firepit. Everything is broken all the time.
