Handle statistics more efficiently
From the Erlang shell, we have
dpm:get_jobs_info/0 to get some insight to what's going on in the DPM. These functions use a brute-force, query-every-process approach, which can be fairly expensive when there are thousands of requests. This issue proposes we use a different approach.
A process can be launched which accepts messages describing changes in internal state. This process can keep accumulated results so asking for these reports would be very fast. So fast we can entertain the idea that a web page could should you the changing, internal statistics of all the DPMs!
The process would keep track of:
- number of ACNET clients
- number of DI/PI/FTD combinations (and how many clients asked for the same set)
- number of FE/FTD processes