Below is the scenario:

  1. I'm running jmx script using non-gui mode.
  2. 1st iteration results are getting high average response (7 secs) with highest latency(1.5 secs).
  3. 2nd, 3rd, & 4th iteration results are giving less average response (3.5 secs) with low latency (400 ms). As latency reduces average response time reduced.
  4. If I restart my server and run same jmx file again repeating steps 2 & 3.

Q1: Is it really a problem with Network Latency. If so how to handle this?
Q2: Is it anyway related to Cache then I have used DNS Cache Manager, Cookie Manager, Cache Manager? Is there anything I forgot?

  • Just to be sure we're talking about the same thing, how do you define the "latency"? How do you measure it?
    – Alexey R.
    Commented Mar 19, 2019 at 11:11
  • How much time the request and response spent on the wire excluding server process time on a particular request.
    – hareesh
    Commented Mar 19, 2019 at 11:28

1 Answer 1


Well, if to refer to JMeter documentation

JMeter measures the latency from just before sending the request to just after the first response has been received. Thus the time includes all the processing needed to assemble the request as well as assembling the first part of the response, which in general will be longer than one byte

So that means that latency time includes the time the the server takes to prepare data for response. What you actually are describing resembles the case when the application under tests needs some time (or some load) to warm up (see the example here) and get to the most effective performance values.

This explains why you're getting back to higher values after restart.

  • Thank you. We will monitor next time. What could be the ideal warmup time.
    – hareesh
    Commented Mar 20, 2019 at 10:01
  • Ideal warm up time is zero. Which means no warm up is required. However this is hardly achievable. You should go from the business requirements regarding application performance. In production you can warm up your application as the part of deployment with generating corresponding load not waiting for the real user queries. However this is not always gives benefit. See details here: section.io/blog/what-is-cache-warming
    – Alexey R.
    Commented Mar 21, 2019 at 11:55
  • Thank you for taking time to respond. Apart from warmup, Are there any other recommendations to take care for actual scenario I faced.?
    – hareesh
    Commented Mar 22, 2019 at 12:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.