benchmarking heavily depends on the size of your cluster and its node sizes, as well as the datastore used and your schema