3 Replies Latest reply on Nov 29, 2017 9:56 PM by Justin D'Cruze Branched from an earlier discussion.

    Performance of views over time assistance


      Hi Gary ,


      Do you any reporting out of Postgres ?


      I am looking to create a view which shows performance of Views over a period of time. But currently Postgres data is available only for 2 weeks.


      Thank you !

        • 1. Re: Performance of views over time assistance
          Matt Coles

          I've authored some workbooks which contain data sources pointing to the PostgreSQL repository database on Tableau Server to allow the general public to benefit from them.


          For general workbook/view/data source information and stats, including Description, check out TS Content. It's got raw user access stats, unique user access stats, content size, owner information, everything you could ever want!


          Satish Parvathaneni, unfortunately, the Tableau Server repository does not contain performance information for viz loads except in the http_requests table, which is actually only retained for 7 days when tabadmin cleanup is run. Audit information on who accessed what, when can be explored with the TS Events data source. I keep asking for it, but no luck so far. You could also parse the text log files to extract some of that information. Easier still if you've got them ingested into Splunk.

          • 2. Re: Performance of views over time assistance

            Thank you Matt. Attached is what i have which show the performance of views over a period of 2 weeks. But i am looking to having for 1 year,as we know we don't have that.


            So i am trying to load the data to a different table in Postgres ,using an ETL tool so that the information is captured daily.Will need to check how that shapes up.

            • 3. Re: Performance of views over time assistance
              Justin D'Cruze

              I like Matt's Splunk suggestion if you have that as an option.


              2 other things you could do:


              1. Run ziplogs reqularly and archive them for later use with Logshark to check performance if needed
              2. Run an incremental extract refresh of the http_requests table (and hope you don't accidentally run a full refresh!)