6 Replies Latest reply: Feb 20, 2018 7:08 AM by Miguel Angel Baeyens de Arce RSS

    Qlik Sense App, QVD & Server Memory Estimation

    Alip Nath

      Hello Folks of Qlik Community!

       

      I am looking for inputs to gather ideas on how Qlik Sense Server sizing and future upgrades to be made.

       

      Consider the situation that your organization have the Qlik Sense Environment set up.

      With time, user adoption has gone up exponentially and there are performance issues seen including outages.

       

      What are the key factors that needs to be looked at and how to do they link to server aspects such as RAM & Disk Space.

      Additional question on these lines:

       

      • Is there any limit to QVD size that needs to be enforced?
      • Is there any limit to Apps size that needs to be enforced?
      • Is there any way to track intuitively track bad performing Apps via kpis like long running loads, high memory consuming apps?

       

      Have read posts on the community but I am unable to find information which is concrete.

       

       

      Thanks,

      Alip

        • Re: Qlik Sense App, QVD & Server Memory Estimation
          Santosh Reddy M

          Hi Alip,

           

          What is the Ram and Hard Disk size for your current environment.

            • Re: Qlik Sense App, QVD & Server Memory Estimation
              Alip Nath

              Hello Santosh - The Server RAM & HDD were set up initially with a high level understanding of usage.

              Currently it is quite higher as I mentioned above; hence we are looking at inputs what should be considered for Server Sizing going forward per the points above.

               

              Thanks,

              A

                • Re: Qlik Sense App, QVD & Server Memory Estimation
                  Miguel Angel Baeyens de Arce

                  To questions 1 and 2, the answer is no: as long as there are resources, the size of apps and QVDs does not affect. Instead, a combination of factors will do: lots of users in a big application with a complex data model and expression in the UI.

                   

                  The QMC and Governance Dashboard or similar applications will tell you how the resources are being utilized. For a fine tuning, use Windows Performance Monitor

                   

                  Qlik is not only about amount of RAM, it's also about speed of RAM, power management of CPUs not just more CPUs, disk I/O, network bandwidth but also latency... Are you virtualizing/using cloud services like Azure or AWS? What is a "long running load" (which you can check in the QMC)? Are you using a multinode and if yes, using shared persistence or syncronization? Do your servers and clients have GPUs?

                   

                  But more important: how your apps data models and expressions look like?

                   

                  Without more details we cannot tell more than follow the Qlik Scalability group for resources and general best practices.

                    • Re: Qlik Sense App, QVD & Server Memory Estimation
                      Alip Nath

                      Thanks for your response Miguel.

                       

                      We have a single node architecture except for the Dev Environment, where we have a separate reload node.

                      By long running load I mean manual load run by users from their work-stream which at times have been running in the background for long hours. Thus bringing down the server as well.

                       

                      Best practices have been followed wherever applicable including App development.

                      What I am looking for is guidelines to use to review the current state of affairs(QVDs,Apps,Loads,Server Resources,Users volume,etc) and check what needs to be improved or scaled.

                        • Re: Qlik Sense App, QVD & Server Memory Estimation
                          Miguel Angel Baeyens de Arce

                          One starting point, if you think it could be hardware related, it could be using the Scalability Tools on the existing applications to see which one is really consuming more resources. With these tools you can simulate how the system behaves with one single user or any number of users, and use this to determine the breaking point.

                           

                          Once you have this identified, if not already, you should start checking the data model and the expressions using in the front end.

                           

                          If you cannot find any room for improvement, then either the demand of more users requires new, more powerful hardware as everything is working as expected, or you will need to engage with Qlik Consulting so they can tell you, for your specific applications in your specific environment, which actions you should follow.

                  • Re: Qlik Sense App, QVD & Server Memory Estimation
                    Carlos Londono

                    Alip,

                     

                    Check the Qlik Sense scalability tool:

                     

                    Qlik Sense Scalability Tools (Sense v3.x, June/September/November 2017)

                     

                    I hope this helps!

                     

                    Thanks,

                    Carlos