I have a question, if someone can help me.
I am trying to run some performance tests using LoadUI PRO test tool.
The licence I have supports 1000 VUs.
So, for the performance test, I am using the following setup:
- Duration: 10 min
- Load Type: Rate
- Peak arriving users: 40/s (from my knowledge, I am sending 40 messages per second which means 2400 per min, right?)
The problem I have is that the test is starting properly, and after aprox 1 min, I receive the message:
"Your licence allows simulating up to 1000 virtual users. To simulate more users, upgrade your license. Your test will continue, but LoadUI will queue incoming requests if the limit is reached".
Also, something else that I have noticed is that during the 1st min I started the test, I didn't noticed that more than 100 VUs (out of 1000) are running --> Running: 95/1000. (Attached you can find also a printscreen).
Can someone help me to understand also if load tests are re-using the virtual users during the run?
Any feedback is kindly appreciated.
With your test configuration, the expected behavior is as follows.
1. During the first minute, the number of arriving VUs each second *will gradually increase from 0 to 40 VUs/s*. Here are the approx. numbers of arriving VUs at some seconds to give you an idea of the behavior:
* 1st second - 0 VU is started
* 12th second - 8 VUs are started
* 24th second - 16 VUs are started
* 40th second - 32 VUs are started
* 60th second - 40 VUs are started
Thus, during the first minute about 1200 VUs are expected to arrive (40*60*1 / 2).
2. If we assume that it takes each VU quite long to complete the scenario, the 1000 concurrent VUs limit may be achieved less then in a minute after you start your test.
3. Starting from the second minute until the end of the test, the number of VUs arriving each second will be 40VUs/s. This means that about 21 600 VUs are expected to arrive (40*60*9).
The "upgrade your license" message in your screenshot looks expected, while the number of running VUs doesn't look expected (see the attached screenshot). Please submit a ticket so that we could look into this more closely: https://support.smartbear.com/message/?prod=ReadyAPI
From the results of that 1 min while the test is running I noticed that the response time of the requests is somewhere around 100 ms, and my impession was that the sockets are being closed and VUs are re-used during the same load test.
But from what I can see, this is not happening. Can you please confirm?
Also, I started to configure some agents (so LoadUI to send the requests to the agents). Do you think this might help me in executing a full performance test (10 min with 40/s)?
Thank you very much for your help.
Have a nice day,