Operator should query K8S API to see if a pod is terminating before attempting to add back

Because when a pod has been sigtermed, it can report itself as the same status as if it needs to be added back to the cluster, we should check with the k8s API and see if there is a terminating status. If so, we should skip the add back.

Additionally, if we do not have a terminating status on the pod, we should allow 30 seconds (3 reconcile loops) in case the termination was facilitated outside of K8S control.