In the course text at point 2.15 there is this sentence:
"In the cluster this daemon monitors both the kubelet and docker processes." This sounds as if Supervisord runs centrally in the cluster - is this the case? Or does it run on every node?
It runs on the nodes that will execute workloads (i.e. the nodes that have a kubelet/docker installed). It's just here to ensure that docker/kubelet get restarted in case they stop.
Also, from what I can see the in code of Kubernetes on Github, they now use systemd instead of supervisord.