3
votes

I am running performance tests on an API endpoint, and I've noticed that when I use the GUI version of JMeter, I get lower response times than when I run the tests in non-GUI mode. For example, a simple request with 10 threads will generate the following:

GUI:
Samples (threads): 10
Min: 765
Max: 853

NON-GUI
Samples (threads): 10
Min: 1114
Max: 1226

I've repeated the requests several times, but the results are consistent even with higher numbers of threads (100-150). I have disabled all other threads/listeners for the non-gui mode, and in GUI mode I only use Aggregate report. Did I miss something?

1
Hello, Any feedback on answer ? If answer is ok, you should accept it and upvote. ThanksUBIK LOAD PACK

1 Answers

0
votes

There are following possible explanations:

  • In Non GUI mode, injection is not parasitized by handling of GUI by Awt-thread or listeners management. So JMeter would be injecting correctly expected load and leading to degraded response times from your server

  • Your are load testing from a different machine in non gui mode, in this case check you are really in same configuration (jvm, jmeter, machine, network)