1 Reply Latest reply on Mar 31, 2016 7:45 AM by pathduck

    JAVA HEAP MEMORY

    karthikraj
    karthikraj Mar 31, 2016 7:28 AM

      Hello All,


      I have decided to monitor maximum heap usage and used heap of my EAP applications (using rhq/JON3.3)
      Pulled out both parameters for SINGLE node and it showing as max heap : 682.7MB and used heap to be 363.2MB

       

      Also  have pulled out the same for all the nodes in my environment and it showing as max heap as 1.8G and used heap to be 1.4GB

       

      My environment has more than 20 nodes,I checked the individual nodes and observed their heap usage to be more than 300 MB (each) so I concluded that Group metrics (second pic) displaying average(not aggregate) heap usage of all the nodes.
      If my conclusion is wrong what does the values in second pic means?If I am right how to view aggregate (not average) heap usage of all the nodes..

       

      If you  are not getting what I am trying to say please let me know I will try much better to post my question clearly.Thanks in advance

        • 1. Re: JAVA HEAP MEMORY
          pathduck
          pathduck Mar 31, 2016 7:45 AM (in response to karthikraj)

          What are the Xmx settings for your servers? Are they all the same?

          What are the Xms (inital heap) values?

           

          I think that like you say, Group Metrics are an average of the values over the time interval you have chosen for this dashboard. Not really sure on the details, I've been confused on this point as well.

           

          Note that Maximum Heap Usage is defined as "The maximum amount of memory that can be used for memory management.". I would assume that since Java will expand the allocated memory (up to the Xmx value), you will see different values dependent on how much memory load each server has. So you might see some servers using 600 and some 1.8GB.

           

          Note also that expansion of heap is a CPU-expensive operation, so you might want to ensure all your servers have the same amount of load to spread it out.

           

          For instance, here are the used heap values from a group of six servers in production. The Used Heap value in the dashboard portlet was 4.4GB so I assume this is the Average value.

           

            • Actions