Which of these events will cause the kube-scheduler to assign a Pod to a node?

If a Pod crashes, the kubelet or a controller typically handles restarts or replacements, but the kube-scheduler is only involved if a *new* Pod needs to be scheduled.

When a new node is added, the kube-scheduler does not automatically re-assign existing Pods; it makes that node available for *new* Pods to be scheduled.

High CPU load on a node typically does not trigger the kube-scheduler to re-assign *running* Pods; existing Pods will continue, potentially with performance issues, unless other autoscaling or rescheduling mechanisms are in place.

The kube-scheduler is responsible for watching for newly created Pods that have no node assigned. When it finds such a Pod, it selects the best node for it to run on, considering various factors like resource requirements, policies, and affinity/anti-affinity rules.