At 13:00 UTC we've marked this as fully resolved.
At 12:20 UTC we detected 502s on Swiss servers that may have been related to a Kubernetes version change. We have made a fix and service is restored, but we are still monitoring the load balancer and Kubernetes cluster.
