We are truly sorry for the interruption in service. After initial analysis we have learned that the connection error was caused by an attempt to elevate a change in production that had to be rolled back.
We had carefully tested the same configuration change in our test environments and considered this a low risk update without any disruptions in service. Unfortunately the update had a different behaviour in our production environment and had to be rolled back. We are still conducting analysis internally to further understand the difference in behaviour. We will ensure to schedule maintenance before attempting the same update in the future.
Thanks for your patience and understanding.