2 Replies Latest reply on Jan 8, 2018 5:50 AM by Kalyan Allam

    Loading time difference in Tableau server with multiple datasources

    Kalyan Allam

      Hi All,


      Recently I am working more on the Tableau server side and so my focused shifted a lot into data extractions and optimizing Tableau dashboards for better performance. Here is my issue -


      We have a large Tableau workbook with 11 dashboards and for some unavoidable reasons we cant break the dashboards into several workbooks. And we are using a live connection to a tde file in Tableau server. There is a landing page in the workbook that visualizes high level data and from that dashboard we have links to other dashboards. Currently it takes 30 to 40 seconds to load the landing page. ( the time it takes from clicking the dashboard to see the dashboard open in tableau server). We want to decrese this load time.


      One of the things I did was i did a extraction instead of using the live datasource and uploaded that workbook. The size of that file in the server is 430MB (the huge is due to optimizing the fields and also I checked the Aggregete to the visible dimensions option in the extract window). Now the load time came to 18 seconds. It is a huge improvement. But we wanted to bring it down to below 10 seconds.


      My next idea was to do a seperate extraction for the landing page and use two seperate data sources in the main workbook. One for the landing page and the other for the rest of the dashboards. The datasource for the landing page was 44 mb (on the computer). I did that and uploaded the workbook. Here is the kicker. Now the loading time is 20 sec or more. I dont get it. how can it be more ? When I publish the landing page as a separate workbook, its loading time is 12 secs. Why is this more when i include this landing page with the other 10 dashboards in a single workbook ? Shouldnt the landing page only load its datasource when i open it ?


      It would be great if some one can shed some light on how this works.


      Thanks every one.