A bug in our application caused any request to connect to our instances fail. The specific issue has been resolved, and a long term solution is being worked on.