0

For example, there is a thread group named login.

Now this login thread group has 10 request where all requests are parameterized(One API request dependent on other) and time b/w request is 1 secs(Used constant timer)

My question is what is the required ramp up second for thread group?

According to actual calculation read from various sources.If i give thread count 50 and ramp-up seconds as 50 then the time b/w thread is 1 seconds where 50 users will be created in 50 seconds but if i use this scenario how the time b/w request inside the thread will be handled.To run 10 request at least jmeter need 10 seconds since i have used timer

If my scenario is correct then the actual rampup seconds will be dependent on number of request and time b/w request.

Then the Ramp-up-seconds should by calculated as below

Ramp-up-seconds=product of(No of requests,time b/w requests inside thread,No of thread count)

Please make some review on my scenario and provide a answer. If i work with actual Ramp-up-second calculation API paramaratization doesn't works properly and my jmeter load test script gets failed

Jmeter Test script will be like as below

  1. Make login: Will perform login and extract the authorization token for further requests
  2. create User: Will created user and createduserid will be extracted from response
  3. Fetch User Detail: Will fetch the userdetail for created userid
  4. Logout: Will perform logout

So this should be happen in sequence

and the time b/w request inside thread is 1 seconds used constant timer

5
  • This question same as [stackoverflow.com/questions/59711844/…
    – Meet
    Commented Feb 4, 2020 at 12:57
  • @Meet It is raised by me as i didn't received proper response there i have posted in QA stack exchange Commented Feb 4, 2020 at 13:06
  • sorry @Mohamed Sulimaan Sheriff but it is raised by zeeshan in StackOverflow.
    – Meet
    Commented Feb 4, 2020 at 13:13
  • @Meet - zeeshan is the OP's Stack Overflow account name. There's no impropriety, it just looks odd.
    – Kate Paulk
    Commented Feb 4, 2020 at 14:21
  • Hey now i have deleted the question in stackoverflow Sorry for this confusion this happened because i didn't got clear answer for my question in stackoverflow that's the reason i moved here Commented Feb 5, 2020 at 4:29

1 Answer 1

0
  1. Each JMeter thread (virtual user) is executing Samplers upside down, in your case sequentially.
  2. Ramp-up doesn't have any impact either on other threads or on sequence of requests execution

The main idea of the ramp-up (and eventually ramp-down) is that it's better to increase the load gradually, this way you will be able to correlate increasing load with other performance metrics (throughput, response time, number of errors, etc)

If your script fails the reasons could be in:

  1. Your correlation logic fails somewhere somehow, inspect variables you're passing between requests using Debug Sampler and View Results Tree listener combination
  2. Your application gets overloaded hence fails to provide normal responses, check response body for failed requests and for Samplers immediately previous to failed requests, application logs, whether the application has enough headroom in terms of CPU, RAM, etc. using JMeter PerfMon Plugin
2
  • Application works fine but my Jmeter doesn't run the URL in sequence i need to run it in sequence if i provide more sampler time it works fine for lesser time parameterization doesn't occurs because of the timing. Only if it run in sequence i can run test Commented Feb 4, 2020 at 11:40
  • It does run requests in sequence for each virtual user, you can double check yourself by adding __threadNum() function to report the number of current thread (virtual user) and __groovy() function ${__groovy(vars.getIteration(),)} to report the number of thread group iteration to your request label
    – Dmitri T
    Commented Feb 4, 2020 at 11:44

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.