Forum Discussion

johny2010_1's avatar
johny2010_1
Contributor
14 years ago

Run Time usage problems

Since run time graph results not written to a file , how can we benefit from them in best suitable way ?

I can see some numbers during test run,



for example CPU usage for a sever,



so, I can get max CPU during a test,

but how will I know that this max cpu usage corresponds to which request ?

I can see CPU usage versus time, can I see CPU usage versus passed request.

But don know which reuqests actually makes high CPU usage.



Ant this issue is not about only CPU but all coutners.

the problem is about timing issues



I see that all counters works due to time, which seen on y axis.



So , these times must be related to actual test times, request times etc.



the problem that we can not see any time in the log results, I mean occuring times for responses.

So, how will I know any counter at any time corresponds to which spesific request ? or connection ?



So, how can I use any of these counters for a result report since I do not know corresponding action, connection or request ?


3 Replies


  • Hello Johny,





    Since run time graph results not written to a file


    We will consider storing the graphs along with the execution results, so they can be used later for analysis.





    I can see CPU usage versus time, can I see CPU usage versus passed request.


    TestComplete does not provide this information - actually, different VUs could execute different requests at the same time moment, taking into account that different tasks can be assigned to different VUs. And the same is true even when the same task is assigned to all VUs.

    You can try to find out what requests were executed at a specific time moment by checking server logs on your web server - there should be timestamps for requests there.
  • Ok,

    Here is my solution to this problem,

    Many requests for CPU at a moment, but if we think reverse of it, just one CPU value for all requests for a moment..



    An adiditional column may be added to requests panel inside Users log which shows CPU value for every request.

    And also more additional performance metrics may be added to requets.



    And that can be arranged as an extra option before running tests if performance degredation occurs, just like request details.

  • Hello Johny,





    Thanks for sharing your thoughts - I have added the suggestion to our DB.