We are investigating this issue, too many requests is made in one second.
Meanwhile we allocated more memory resources during this time temporarily.
Update 09:43 pm — Server has been stabilized. We are still in the midst of finding out what happened.
Update 12:12 am Sunday — It began when one of our servers had an overloaded queue due to a bug introduced in a recent update. This server attempted to resend too quickly to another server which couldn’t keep up. The whole chain reaction is the reason for our 2 hour service disruption. We are optimizing our load right now.
Tags: memory, overload, resource, vertical scaling