0
votes

These days I am trying to use JMeter to do load test. There is a very confusing point I can't understand: The distributed environments: 1. Three 8 CPUs 16G servers running JMeter Server 2. 4 CPUs 8G server running JMeter 3. These servers are on the same subnet.

The Thread Group settings: Threads number: 2000 Ramp-up: 0 Loop: 10 The Throughput is 3000/s

and another Thread Group: Threads number: 2000 Ramp-up: 1 Loop: 10 The Throughput is 5000/s

and another Thread Group: Threads number: 2000 Ramp-up: 1 Loop: 5 The Throughput is 4000/s

So, I have no idea which one is the right Throughput, or should I have to try different number, ramp-up, loop to get the highest Throughput? Or I am wrong, someone help, it's very confusing?

1
Actually Ramp up 0 is not real scenario , 2000 users/requests do not come in 0/1 second if we imagine , you should create some feasible test plan and then you can check everything based on that.Helping Hands

1 Answers

1
votes

To load tests I'd recommend ThoughoutputShaper plugin (from http://jmeter-plugins.org/wiki/ThroughputShapingTimer/).

It provides intuitive interface to experiment with TPS/RPS and is very efficient when it comes to generate required load.