Overpass API developpement

[<>]

Herr Olbricht
    Massive Massive thanks for your time. You're answers provide great
insight.

    In your response you indicate:
However, the dispatcher may delay or refuse to give to a process a share
of runtime because there are already too many other requests of the same
client, or under higher load simply too many active concurrent processes
in total.

   2 questions:
   a. By 'active concurrent processes' you mean active interpreter processes
i.e. those spawned by cgi or do you mean the all the processes running on the
machine.

    b. In my setup. I have two clients from the same machine(i.e. identical
source IP) contacting the overpass server. I wish to disable any significant
pause in overpass that might limit clients from the same IP. I have rate-limit
as 0 on dispatcher start up. I also read on the internet a suggestion by user
mmd, recommending reducing the millisleep time in dispatcher.cc. Will this
minimise the dispatcher's tendency to share the runtime?

Thank you