Quote Originally Posted by the sniper View Post
I think the article explains how it happened pretty well?
Miss-integration between two services is well explained?
At more than 3 million requests per hour, the public-facing API became overloaded
This is something that a lot of services have trouble with after starting out or making major changes and I just can not understand it. How can they let this happen? The amount of requests can't be a surprise for them, they must have some previous data and be able to calculate a safe buffer.

Still, non of the issues that revealed themselves during the launch was a surprise and I'm quite confident it will be stabilized quite quickly.