this enables control over where pods get scheduled in a kube cluster, using nodeAffinity and tolerations