Target MPS/QPS
Number of parallel queries
(LOWER IS BETTER)
Superior Latency: NaNx faster at P99
Duration, mean latency, and worker fairness
(HIGHER IS BETTER)
Execute x more queries with the same hardware