Forum Discussion

jsheph01's avatar
jsheph01
Contributor
6 years ago
Solved

Should I use VUs or Rate?

I am trying to create a ramp up load test and I am not sure if I should be using VUs or Rate. I was looking at the documentation for load type and it says VUs are not restarted for a rate type load test. Will the VUs not restarting effect how may requests per second I can send?

 

This is what I am trying to setup:

Single scenario distributed over 2 agents

Scenario has 2 different requests to the same endpoint.

VU: 1200 (600 per agent)

Max requests per second: 1250

Expected response: 250ms

  • kondasamy's avatar
    kondasamy
    10 years ago

    Not sure if I am correct, I guess there is no such option in SoapUI OS for maintaining centralized endpoint. But, there is a workaround which I can suggest,

     

    • Create a Project level - Custom Property called Endpoint and then map your URL as the value for the parameter.
    • Then copy the below Groovy code in a Groovy Test step and Run it.

     

    def endpoint = '${#Project#Endpoint}'
testRunner.testCase.testSuite.project.testSuites.each
{
    suite ->
    suite.getValue().testCases.each
    {
        tc->
        tc.getValue().testSteps.each
        {
            it->
            if (it.getValue().config.type.equals("restrequest"))
            {
                 it.getValue().getHttpRequest().setEndpoint(endpoint); 
            }
        }
    }
}
log.info("Done")

     

     

    Now all your REST test steps would have the endpoint value set as 

    ${#Project#Endpoint}

    Hope this helps you!

     

    Thanks,

    Samy

1 Reply

Sort By
  • NBorovykh's avatar
    NBorovykh
    Moderator
    6 years ago
    Copy LinkReport

    Hi jsheph01, 

     

    Both "Rate" and "VUs" load types (in combination with the "Ramp Up" load profile) may be used for simulating an increasing load. The option you select and its specific settings depends on the desired throughput, the desired LoadUI behavior when the tested API starts experiencing performance issue under the load and the length of the scenario.

     

    Both load types should allow you achieving the needed number of requests per second, but with the "Rate" type it may be a bit easier to plan and predict as you don't need take into account the idle periods of VUs lifetimes. Let me clarify this.
    * With the "VUs" load type, before a VU is restarted, LoadUI waits for some time (the "Wait time" parameter). During the wait time, the VU will be counted as working, but it won't be running its scenario (i.e. won't be generating any actual load).
    * With the "Rate" load type, a VU will live only while it will be running its scenario (e.g. it won't have any idle periods during its lifetime).