5 Replies Latest reply on Sep 22, 2020 7:36 AM by Dave Nacy

    PIM Dataquality | performance issue

    Alex _ Guru

      Hi all,

       

      the customer have some issue with the performance of dataquality.

      In this moment if I launch 48 dq configuration for 200 variants, the system complete the job in 15 minutes and I think that this is not a good performance.

       

      48 different configuration for 5 rules.

       

      which parameters can I modify to parameterize the execution?

       

      Thanks

      Alex

        • 1. Re: PIM Dataquality | performance issue
          Sai Narasimha Prasad Rebba Guru

          Hi Alex,

           

          I think there are few parameters in plugin_customization.ini file which we can tune and see if those help for bit good performance.

           

          Thank you
          Sai Prasad

          • 2. Re: PIM Dataquality | performance issue
            Ralph81 Guru

            Hi Alex,

             

            as Sai Prasad mentioned there are different parameters in the plugin_customzation.ini which allows a bit tuning.

             

            But also a good point can be how the rules are triggered. For example a BULK operation (import trigger,..) is always faster then single actions (entity changed). The last one I would only use if there is really no other use case for it.

             

            Best Regards

             

            Ralph

            • 3. Re: PIM Dataquality | performance issue
              Alex _ Guru

              Hi ralph,

               

              The dq are triggered manually by users and we cannot do the same process with a post import.

               

              Can you share me some properties with values that i can try?

               

              Pim 8.1.1.05

              • 4. Re: PIM Dataquality | performance issue
                Ralph81 Guru

                Hi Alex,

                 

                you will find some hints in the documentations about that properties. As every Customer and System is different I can not make any recommendations here. Its just a simple testing to see if its getting faster.

                 

                And if your users are using "entity change" triggers, that will never become very fast because the Framework will be loaded for each item and not once for a bulk of items. So when ever there is a chance to push it into the Job Framework, for example Import/Export/Merge or time, use it.

                 

                I think in less of the cases its really needed that the user is getting a response within seconds.

                 

                Best Regards

                 

                Ralph

                • 5. Re: PIM Dataquality | performance issue
                  Dave Nacy Guru

                  There is a keep alive parameter for DQ rules that may prevent the entire framework from having to be loaded for each invocation. See if changing the value to a longer time interval  of this parameter in plugin_customization.ini helps:

                  com.heiler.ppm.dataquality.server/dataquality.executionPool.executorKeepAliveInterval=3600