I have an issue with a non-linear growth of active threads number. It happens at the beginning of the test and becomes linear later. Important: it occurs during distrubuted
