1 Reply Latest reply on Aug 21, 2016 12:17 PM by Yuriy Fal

    Streamlining Dashboard Performance on Live Connection (Faster Load Times)

    Clarence Magsalin

      Hi All,

       

      I'm putting this out there, someone might have better methods in terms of faster load times of Live Connection dashboards. First off, here's a background of what we have as of now:

       

      - Tableau Server/Desktop Version: 9.2.4

      - Source : Google BigQuery Live Connection

      - Number of Table/s : 5 Tables for 5 dashboards

      - Aggregate table size : 12 million rows, 4.2GB

      - Number of charts/graphs per dashboad: 4-5

      - Initial load time in server (uncached) : 30-40sec, 1 dashboard 60sec

      - Initial filtering, uncached (by date/category) : 8-10sec

      - Cached load time in server (refreshing dashboard after initial load) : less than 10sec, sometimes 25-28sec

      - Cached filtering : 3-7sec

      - Data granularity : row level (by Account ID) - we cannot disolve this, as we are collecting "data buckets" based on account ID

      - Prior to import to BigQuery, data is grouped per categories using GROUP BY

      - BigQuery using DAY partition

      - Level Of Detail (LOD) calculations in Tableau were stripped off - LOD is not allowed on Live connections

      - almost all calculations and date-to-string transformations are eliminated from the dashboard - calculation was done in the scripts, using aggregate columns

      - Cannot switch to TDE (Extract) - for security reasons, this embeds user login unencrypted

       

      What we want to achieve is to slash the load times during initial load and avoid "Browser Processing Error" when multiple users are firing up the boards.

      What we're considering are:

      - Increase the Tableau server cache

      - Increase the time the cache keeps the data

       

      I'm not sure yet if those options are possible since our client's admin keeps the server and we have no visibility to it.

      Any new input is really appreciated, since most of the forums, KB and whitepapers I read are pointing to TDE which is really not an option.

       

      Thank you!

      Clarence

        • 1. Re: Streamlining Dashboard Performance on Live Connection (Faster Load Times)
          Yuriy Fal

          Hi Clarence,

           

          12M rows look not so much for BQ,

          so table size is unlikely the problem (imho).

           

          That's why the following questions:

           

          1) How many Marks are rendered on a view / dashboard?

           

          2) Are you using quick filters? How much unique values are? Do any filters have Only relevant values flag set?

           

          3) Is it a chance that you're using data blending? If so, how much unique values are in blended field(s)?

           

          4) Are there any difference in load/render time when in Tableau Desktop as compare to Tableau Server?

           

          5) What steps are the most time-consuming -- based on Performance Recording data?

           

          Other questions may arise regarding to BQ itself,

          but those above are common from Tableau side.

           

          Yours,

          Yuri