Tolerate node down taints for workloads tied to nodes

It makes no sense to have Kubernetes reschedule them, as they are tied to a specific node via affinities anyway.

Merge request reports

Loading