Even with a timeout, the time spent waiting on the timeout is wasted utilization. Therefore, if the request rate stays the same, and each request is wasting enough utilization, the utilization required to process the request will be higher than the rate it is worked off.
