At the moment we are facing an issue with slow etcds which makes some operations impossible, like updates to the cluster. Running workloads should not be affected unless they use the Kubernetes API Server directly.
We will keep you informed and apologise for any inconveniences.
Update: We have recovered from the incident.