Analyze possibility to report polling distribution for concurrency reduction measurements

[mensfeld]

One common misconception about karafka multi-threading is that it works like Sidekiqs. The tricky part is, that it works and it does not work similarly. Aside from explaining, etc we could provide a metric/stats in web-ui per process on their real-time ability to operate in multi-threading. It's not the same as utilization, because utilization takes into consideration work time but here we would be interested only in polling ability/reduction of threading potential.