Forum Discussion

stevenhum's avatar
stevenhum
New Contributor
15 years ago

HTTP Load Testing Disimiliar Results In Return

Dear AutomateQA Experts,



Would wish to know the reason why everytime a HTTP Load Test is being executed, it returns a disimiliar result in return ? For instance, there are 10 users (ranging from A to J). On first attempt, person A to H suceeded, but person I and J fails. On second attempt, person B fails, and the rest of the users suceeds. The more times the load test execution is being undergone, various results differ in the process.



Kindly assist on the matter.



Thank you.

3 Replies


  • Hi Steven,





    A warning in Load Testing results does not necessarily mean that the execution failed.

    The Warning status is assigned to a request when the actual server response code differs from the recorded server response code. What are the recorded and actual response codes in your case for the requests with the Warning status? The recorded response can be found on the Response tab of the LoadTesting Task editor, and the actual code is displayed in the Headers panel to the right of the requests list in the Load Testing Log.
  • stevenhum's avatar
    stevenhum
    New Contributor

    Dear Alex,



    I have encountered 4 disimiliar erros at the headers panel, namely :-



    i)     HTTP 503 Service Unavailable

    ii)    HTTP 500 Internal Server Error



    As for the 3rd and 4th error, it retuns back either a ReturnURL path or an exception handling error. When I enter the program manually, it works fine. None of the above mentioned error appears on screen.



    Though I had placed in delay time executions in between, the error still remains as it is. One thing in particular is errors does not seems to surface itself when a single user load test is being executed. Problem arises when there are more and more users entering the system.



    For your information, I am trying out the login load test that encompasses 50 virtual users. On several attempts, all test executes fine, on other attempts, warnings appears. Am still in doubt over the matter as to why the results show inconsistencies in between.



    Be appreciated if you could enlighten me up on this.



    Thank you.



    Regards,

    Steven Hum

  • Hello Steven,





    Taking into account the fact that one virtual user always works fine, I guess that the virtual users sometimes get overlapped in the sense of session/identification IDs (not sure what exact IDs and where they are - need to see the task and the results to say).

    What I mean is that all users send the same ID in some requests, while the server expects different IDs. To say something more definite and make some suggestions, I need to see the task and the results of its execution - please zip your project suite folder and send me the archive:

    http://www.automatedqa.com/support/message/