problems are dispatched to the same nodes in the cluster to ensure cache reuse where possible