3 Replies Latest reply on Feb 3, 2015 8:11 AM by Jeff Strauss

    Backgrounder.querylimit behavior

    Kevin Doherty

      I had a quick question around the backgrounder.querylimit setting and how it differs in regards to Workbooks with embededded Data Sources versus Server Data Sources.

       

      For Server Data Sources it is straightforward as you are refreshing the one object and once it reaches the querylimit it will fail.

       

      What I am somewhat confused on is a Workbook with multiple embedded Data Sources.

       

      As an example, if I have the backgrounder.querylimit set to 30 minutes and I have 5 embdedded data connections that each take 10 minutes to Refresh. Would this Extract refresh fail?

       

      Does the # of Background processes play into this?

       

      So if I have just ONE Background process running on my server, in the above example, the Workbook would take 50 minutes to refresh due to the single-threaded Backgrounder running one Extract at a time. Would this extract fail if the querylimit was set to 30 minutes? Or would it succeed as each Data Connection only takes 10 minutes to refresh?

       

      If I had 5 Background processes setup and the Workbook was the only object scheduled to be refreshed, the refresh of the Workbook would then take 10 minutes, so would that succeed?

       

      So in the above examples the number of Background processes available plays into this with querylimit and is not as cut and dry when you have a Workbook with multiple embedded connections (Especially as you begin to schedule more and more objects during the same time frames).

       

      Just looking to see if anyone else has input into these scenarios and settings?

        • 1. Re: Backgrounder.querylimit behavior
          Toby Erkson

          Great question!

          I have a 3 hour limit set on ours but I have seen one process go past that limit.  When I asked the user about their extract they said that the report calls some stored procedures and those procedures takes 10-20 minutes to run so I'm wondering if that wait time isn't counted in the limit calculation since the process is possibly idle while waiting for the data?

          • 2. Re: Backgrounder.querylimit behavior
            Kevin Doherty

            I haven't dived into how the DB session activity relates to the Extract session. Not sure when the Extract session starting time would be considered.

             

             

            I was more trying to understand how the querylimit setting ties to a Workbook with multiple embedded data sources. Not sure if it is tied to the Workbook refresh as a whole or to each individual Data Source refresh time within the WB.

             

             

            I have pinged Tableau support and some onsite vendor resources and I don't get any consistent answers even with the controlled scenario I presented above.

             

             

            I am meeting with our Tableau Technical rep and will pose the question to him, I will update this thread with what information I receive.

            • 3. Re: Backgrounder.querylimit behavior
              Jeff Strauss

              great question.  Let us know the definitive answer when you get it.