AW: Scaling the number of users instead of the hit rate

[Hohl, Gerrit]

Hello everyone,

okay, maybe I wrote a little bit too early that my formula is working.
The number of host and robots is correct. But something different is
odd: If I run a test with the configuration beneath, I would expect a
request rate of 500/sec in sum. But I get a request of about 2000/sec at
each of my two polygraph-servers, means about 4000/sec in total. What
did I wrong? Is the peak_req_rate per robot? Or does the max_host_load
have any influence on that? I thought it is only the maximum rate on
which the simulated servers should respond. Should I lower it also to
500/sec? Or does it apply for each simulated server, or maybe for each
polygraph-server? Or maybe there is a bug and it ignores 'min' and still
thinks that is "6/sec"?

Bench theBench = {
        peak_req_rate = 500/sec;

        client_side = {
                max_host_load = 20000/sec;
                max_agent_load = 6/min;
                addr_space = ... ;
                hosts = ... ;
        };

        server_side = {
                max_host_load = client_side.max_host_load;
                max_agent_load = undef();
                addr_space = ... ;
                hosts = ... ;
        };
};

PolyMix4As asPolyMix4 = {
        agents_per_addr = 2;
};

Regards
Gerrit