This morning (December 1st) between 05:44:09 and 05:46:42 UTC the API exhibited longer response times and in some cases returned an error response. The service self-healed and was back to normal operation at 05:46:59 UTC. The root cause was related to memory utilisation which in combination with how services where arranged in the cluster at the time caused a short moment of resource choking on one part of the system. We are sorry for this, and are working on a permanent solution to it. We will further updates if necessary.
