[RESEARCH] Develop a way to measure how system behaves under load

[AlexPykavy]

For now, the system is pretty naive and there are no configurable limits, meaning that under high load it will most probably slow down and even shut down.

I don't want to come up with some limits (e.g. MAX_TASK_NUM) just out of the blue and decided to develop a way to measure how system behaves under the load and use that empirical data to justify the default limits.

[AlexPykavy]

In the beginning, we will just expose /metrics endpoint and add Prometheus scraping the data every second in the local setup.