AllChannel load balancing/robustness experiment

[devos50]

Continuation of the discussion in Tribler/py-ipv8#979.

Future ToDo, after 2022 summer: resurrect AllChannel experiment. The AllChannels content dissemination community test has "unreasonable effectiveness". By plotting 1000 lines into a single plot a healthy experiment or fault shows up. Many bugs have been shows to exist using this approach. Its basically an end-to-end performance analysis, peer neighbourhood bias, and unhealthy overlay in general. AllChannels used 25 figures to depict the code health. One expert glace is sufficient to detect some new introduced bug.

One expert glace is sufficient to detect some new introduced bug.

Preferably, this should be an automatic trigger. I remember that a student worked on combining statistics of the last X runs and then computing the standard deviation of various metrics to detect anomalies.