It is a bad day for some of the Amazon Cloud (see this for a report how Reddit, Airbnb, Github, CourseEra, Flipboard and many more went down).
Fortunately, we have been working on expanding our redundancy and failover. And this was a good test of these expanded capabilities. We were able to survive without manual intervention. Our affected servers failed over to a different Amazon Zone without impacting customers. You can read more about our Redundant Cloud Architecture in this post, as well as see our current architecture diagram here.
