We need a way to gather overall system disks stats to provide in reporting.realtime
This is mostly for TC but maybe down the road to use in our dashboard.
The idea is to use or implement a tool that will get an average between all disks, e.g.
gstat -p -C (I dont know linux equivalent)
disk busy percentage
current read/write rates
We need to keep performance in mind, we dont want to increase the CPU usage too much since this may be running 24/7/365.
One idea is to write a Cython (releasing GIL) or C/Go (pipe) that will read everything, make all the calculations and only then write to a list (Cython) or to pipe (C/Go).
Doing it in plain python may be fine as well but we need to profile it with in mind that we may have up to ~250 disks.
We will need this for SCALE initially but may eventually have to backport to 12.