Background workloads should not have their iteration and thread counts multiplied because they already have a very large iteration count (Number.MAX_SAFE_INTEGER) and because they'll be performing actions that should only be done by one thread, like stepping down a node.