Restarting affected service solved the issue. Cause was an unusual memory behavior and we are investigating further to avoid this in the future.
We see some beta instances failing due to a faulty service. Investigating and working on it.