locked
Visual Studio Load testing issues RRS feed

  • Question

  • Hi All,

    Has anyone attempted >4K concurrent users load test using Visual Studio? I would like to understand the best practices and how they handled the agent overload issues.

    I start getting 'SocketException'  An existing connection was forcibly closed by the remote host.I get several thousands of such errors during less than 10 mins of errors.

    I am unable to understand, if this issue is coming from the agents or from web server? The event viewer on controller and agent doesn't have anything related to these errors.

    I have 200 users running per agent and have 20 A4 roles (Azure VM machines). A4 are eight core/sixteen GB machines..which I think is more than enough to handle 200 users.

    Can someone please let me know, if the Test Agent hosted in Azure has any limitations on the number of connections/Total requests??

    FYI, if I reduce the pool size (Run time setting) to 150, I don't see these socket exception errors. But, this option is not suitable for us as we are expecting total number of requests in IIS to meet as per our requirements. The avg response time is less than 1 second and doubt if there any setting in Azure, which is rejecting our requests? Both app, test controller and test agents are currently hosted on Azure.

    If MS product team or R&D wants to discuss, I am more than happy to connect to customer's environment and show the issue in detail. thanks!!

    Thursday, November 19, 2015 9:44 AM

Answers

  • Hello Jack,

    Thanks for keeping this thread open.

    We see 99% reduction in socket exception errors now and our IT team made changes to overcome these errors.

    I believe they used Azure traffic manager to spread the traffic across two cloud services.

    Thanks!!

    Merry X'Mas and Happy new year!!

    • Marked as answer by Shiv_p Thursday, December 24, 2015 5:01 AM
    Thursday, December 24, 2015 5:00 AM

All replies

  • Hi

    >>Can someone please let me know, if the Test Agent hosted in Azure has any limitations on the number of connections/Total requests??

    I just get some information from the two blogs:

    http://blogs.msdn.com/b/visualstudioalm/archive/2014/01/13/hosting-testcontroller-and-testagents-on-azure.aspx

    http://blogs.msdn.com/b/allendm/archive/2015/07/20/remote-test-execution-using-team-foundation-server-2015-rc2-and-beyond.aspx

    But it seems that no connections or requests limitations.

    This document shares us the Test Controller and Test Agent Requirements for Load Testing:

    https://msdn.microsoft.com/en-us/library/ff937706.aspx

    >>am unable to understand, if this issue is coming from the agents or from web server? The event viewer on controller and agent doesn't have anything related to these errors.

    Not really test it using remote machine, if you use so many v-users in your local running, whether it has the Exception in your local Event viewer? You could enable the Test controller and Test Agent log, maybe the exception was captured by the specific log file:

    http://blogs.msdn.com/b/aseemb/archive/2009/11/28/how-to-enable-test-controller-logs.aspx

    http://blogs.msdn.com/b/aseemb/archive/2010/03/07/how-to-enable-test-agent-logs.aspx

    Best Regards,

    Jack


    We are trying to better understand customer views on social support experience, so your participation in this interview project would be greatly appreciated if you have time. Thanks for helping make community forums a great place.
    Click HERE to participate the survey.

    Sunday, November 22, 2015 7:36 AM
  • Hello Jack,

    Thanks for your response. Appreciate your time in sharing the links.

    My team is in touch with Azure experts and VS product team in Microsoft, as the issue is not resolved.

    I will come back and update this thread, as soon we find out solution to this issue. Thanks!!

    Wednesday, November 25, 2015 5:43 AM
  • Thanks for your friendly response, Shiv_p!

    >>My team is in touch with Azure experts and VS product team in Microsoft, as the issue is not resolved.

    If so, I will keep opening this case for you, of course, if you get any latest information or you get the solution from the product team experts, please feel free to share the solution here:)

    Have a nice day,

    Jack


    We are trying to better understand customer views on social support experience, so your participation in this interview project would be greatly appreciated if you have time. Thanks for helping make community forums a great place.
    Click HERE to participate the survey.

    Wednesday, November 25, 2015 8:40 AM
  • Hello Jack,

    Thanks for keeping this thread open.

    We see 99% reduction in socket exception errors now and our IT team made changes to overcome these errors.

    I believe they used Azure traffic manager to spread the traffic across two cloud services.

    Thanks!!

    Merry X'Mas and Happy new year!!

    • Marked as answer by Shiv_p Thursday, December 24, 2015 5:01 AM
    Thursday, December 24, 2015 5:00 AM