I am testing a node application which handles requests asynchronously. It takes around 5 seconds to respond (or more depending upon the load) for one request.
I want to configure Jmeter with 5000 concurrent threads, and fire 100 requests per second (or lesser). I tried using Constant Throughput Timer, but it does not give me desired results.
What is the correct way to implement 'n' number of requests per seconds from JMETER?
Constant Throughput Timer should be good choice, however it might be the case you're not supplying enough threads in case if your application response time is too high or JMeter fails to kick off enough virtual users.
Another reason is that Constant Throughput Timer is precise enough on minute level, if your test doesn't last long enough it might fail to limit the throughput to the desired value, if this is the case - you will have to play with ramp-up period under the Thread Group
As an alternative you can try Throughput Shaping Timer in conjunction with the Concurrency Thread Group, they can be connected using Feedback Function so JMeter would be able to kick off extra threads if the current amount is not enough to reach/maintain the current load.
