3 Replies Latest reply on Dec 16, 2013 11:03 PM by Chris Marcotte

    What is idle time in Coradient/EUEM?

      Share This:

      Hi All.  I need to understand better what the latency component idle time reported by Coradient really means.  Doco says the time the browser is not making requests ie running java scripts to work out what to fetch next.

       

      However I am seeing substantial variability in how much idle time there is for pages.  There is more idle time for busier periods.  What I don't get is why idle time would vary?  I'd assume that idle time depends on the client computer's capability and that with sufficient clients it would average out.

       

      So I don't get why I see idle time increasing in busier periods and decreasing in less busy periods.  Can somebody help me out??

        • 1. Re: What is idle time in Coradient/EUEM?

          Hello Ric,

           

          Idle time is a metric EUEM records during the page load.  It is a timer that starts when an object is completed (ACK seen from last package of response to a request) and this timer stops when EUEM sees the beginning of the next request.

           

          There are many reasons for idle time.

          1)  Java script loaded from an object and customer's browser is busy executing.  (Browser will not request more objects until Java is complete)

           

          2)  Some objects might be requested from another server that EUEM does not see ie. static content server and traffic going to this server is not seen by EUEM.

           

          3)  Any other object that will cause a customer's browser to stop requesting more objects.

           

          4) Idle time might increase when objects are refreshed from inside the page ie a webservice call.  Page is complete and page timeout is 20 seconds - but a webservice call requests that a jpg on the page be refreshed at 19 seconds.  You would see a big part of E2E time due to Idle time.

           

          If you drill down into one of those pages, can you see where the idle time 'chunks' are occurring?

           

          Hope this info helps.

           

          -Chris

          • 2. Re: What is idle time in Coradient/EUEM?

            Hi Chris.  That does help quite a lot in understanding the meaning of it.  I like to echo back my understanding of what someone tells me so they can can see if I have got it.  EUEM will see and report as idle time any period after an ACK for the last outstanding package till the next request of a package.  Now I think this excludes user think time but not sure how.  I am presuming that a request for a new page or a refresh demarks user think time - somehow.

             

            Some causes of idle time are going to be constants for a given client computer.  Running java scripts, periodic refreshes.

             

            However EUEM will not see time spent outside of its watchpoint.  So if the the client is going off to another url then that time will be seen as idle for a given EUEM but is actually real waiting time for the client. 

             

            For me to see variable idle time that seems a likely area to look into.

             

            I can see what is happening with Fiddler and see 'white space'.  So I guess it is a matter of figuring out how much time is spent going to other urls and local client processing.  IDK if there are better tools on the client side.

             

            Thanks a lot for your help.  I do appreciate the support BMC gives and the folk who give it. 

            • 3. Re: What is idle time in Coradient/EUEM?

              Hello Ric,

               

              Think time is the time between a completed page (all objects loaded) until the next request for a new page.  This could be due to anything (getting a cup of coffee, answering the door) after the last object on the page was loaded.

               

              I'm glad the information helped!

               

              -Chris