Facebook says now that it was a server configuration error. That's easy to do. I've done it. Probably every other person who has ever changed any setting on any server has done it at some time(s). The question here is how such a colossal service (like Facebook or Amazon, etc.) could allow there to be a single point of failure. If they have a single "Authoritative" server that propagates its settings to a bazillion other, dependent servers, then that is a potential single point of failure. Who's the Bozo who allowed that arrangement to exist?
I've brought down a small, client business for 15 minutes or so--two different times, by making required changes to server settings. Oops! And then you fix it. The response from the owner was, "Thanks, Bob." But shutting down a millions of dollars per hour (maybe per minutes), world-wide enterprise for the better part of a day--that shouldn't be possible. There should not be a single point of failure. It shows that Facebook has simply scaled up their rinkydink, dorm room approach.
Bob