Under a heavy load of low-priority requests, it may happen that the processing time of high-priority requests increases due to the delay in freeing up the slots consumed by these low-priority requests.
The goal of this feature is to define a number of slots reserved for high-priority requests. By default, this feature would be set to 0.