As somebody who does DevOps consulting I was baffled how they could be down all day if it was a pure operations issue. That tweet is just speculation though. Hopefully more facts and details come out.

My bet is still on it being a database issue. If they indeed are using AWS RDS, I wouldn’t be surprised if they have half of AWS support engineers working on it. Once again, if I had to bet, I say they rolled their own "half-baked" database solution instead of using a AWS native offering.