3 Replies Latest reply on Feb 8, 2018 7:43 AM by Matt Coles

    Published Data Source Workload

    Brant Davis

      I was wondering if anyone knew where the work was offloaded when connecting live to a Published Tableau Server Data Source (which uses an extract). Will the Data Engine on the server or the local machine take the hit when querying data in order to create visuals. This boils down to on whether or not a single user in Desktop could make a large request using a published data source, in turn hogging resources from others on Tableau Server.

        • 1. Re: Published Data Source Workload
          Matt Coles

          Tableau Desktop will issue a query to Tableau Server, and Tableau Server's Data Server process will receive the query and call Data Engine to execute it. The results will be returned by Data Server to the Desktop client, which will then visualize the results. So the workload will be on both Desktop and Server, but more on the Server side than Desktop. So a single user in Desktop could make a large request to a published data source, and "hog" resources from other users, yes. Though if they are using Tableau Desktop, and the data source is backed by an extract, I feel like they'd really have to be trying hard to impact the Server to really have much of one. It's a bigger deal when users set up automation on their machines to render vizzes every thirty seconds, or they schedule an extract refresh to run every hour.

          • 2. Re: Published Data Source Workload
            Brant Davis

            This is very helpful. Would it be safe to say the Server resource strain would be the Data Engine, and the VizQL process would be the resource strain on Tableau Desktop?

            • 3. Re: Published Data Source Workload
              Matt Coles

              Yes, exactly. But with Desktop, the actual process name will still just be "tableau".