investigation: Does the timer concept overwhelm its own if CPU bottlenecked ?

[SurfingNerd]

It has been observed that on Testsetups in Idle with slow CPUs, the CPU usage stays absurdly high.
the theory behind it is, that the timers even trigger new work, even the old work has not been done yet.

Todo:

• verify this thesis
• -> develop a timer system, that ensures that the timers act as a Timespan between working phases, instead of overcommiting.
  example: in a scenario where a timer is configured to trigger all 1000 ms, and the work would take 2000ms, the timer would trigger in 3000ms, instead of running in an infinite loop, where the system just starts over and over doing what it has just finished. Therefore there would be enough ressources left on the system to maintain TCP Connections.

[SurfingNerd]

another reason could be that the compute heavy HBBFT Messages are running on the same workers than the Session handling.
Therefore instead of handling the HBBFT Tasks on those workers, we could simply just forward the task to new workers pool that do have there own threads. This frees up the worker slots for the communication.