In jmeter, reference m concurrency, you probably immediately think of the number of threads is set to m;
RPS pressure is adjusted server, i.e., the number of requests per second; RPS the number of threads, number of cycles, ramp-up time adjustment;
problem:
There is a page, you need to test the maximum number of concurrent users?
At this time is to be calculated maximum number of users concurrent, simultaneous operation is emphasized, as may be appreciated that while the initiation request;
a) using @ GC JP - Shaping the Timer Throughput (throughput shaped timer, timer adjustment rps);
Setting the number of threads of a thread group 50, ramp-up time of 1s, loop forever; RPS of 1 to 400; the final run time fetching thread group run time and the minimum time timer;
Run to view the results, it was found to increase after the 140 HPS, stable ten seconds, followed by HPS average about 100;
The HTTP request is deployed on the project himself Ali cloud server;
TPS 48 to stabilize the rising ten seconds, and then an average of about 30 TPS;
The response times of less than 3S;
Stable situation, RPS of about 140, a response time of 1.6s, the maximum concurrent = 140 * 224 = 1.6; can be appreciated that satisfies 140 / rps pressure within 1s s to just start threads 224;
To the number of threads in a thread group 100, the other will be unchanged; RPS and tps is substantially the same;
b) a stepped pressurizing thread group
=================================================================================================================================
