# -*- mode: snippet -*-
# name: Workbook performance questions - DE
# key: xwpq
# --
Was the workbook/dashboard re-created according to our best practices and does the problem still persist afterwards?

Please contact the workbook creator and ask him what data sources are used (live or extracts, and what is the underlying data source of the live connections) for the problematic dashboard. We need a clear answer on this in order to continue the investigation.

Please ask the workbook creator what is the difference between the views that are working and those that aren't.  Since we are dealing with a single workbook and the custom views that are and aren't working are known, this should be straightforward. If excessive filters and calculations are used in those affected then we most likely have the root cause of this behaviour.

Please install Tableau Desktop directly on the Tableau Server machine. Try to open the affected custom views in Tableau Desktop directly on the Tableau Server Windows machine and inform us about the results. 

I understand that a live connection to BW HANA is used for the dashboard. Can you please create and extract of this data source and replace the live connection with it. We want to see if the live data source connection might be the culprit and by excluding all live connections we can confirm this if the problem persists with only extracts.

If the problem can be reproduced with only extracts, then please send us a sample workbook with the extract included with which we can test and reproduce the issue on our side. This way I can also open a defect with our developers if needed. If the data is confidential, a sample packaged workbook with sample data that mimics the structure of the original workbook is sufficient. Instructions for creating and sending a packaged workbook can be found at the following link:
http://kb.tableau.com/articles/knowledgebase/sending-packaged-workbook
If the data source is a cube, creating a packaged workbook will not be possible. 


Does the problem appear in a single custom view of that workbook only or for all custom views in that workbook? 

If the problem appears in a single workbook only, then what is the difference between this specific workbook compared to others where custom views are working? (what data sources does the problematic workbook use: live, extract- if live what is the underlying data source, are any specific joins, calculations, filters implemented)? Please also provide the name of that workbook/custom view.

Does the problem appear for all users the same, including the Tableau Server administrator? 

Please confirm, does the problem appear in one specific workbook/workbooks or does it appear with all workbooks published to Tableau Server, also with the sample workbook?  Also describe the complexity of the workbooks in question.
 
What do the problematic views have in common in comparison to the dashboards which are loading as expected?
 
What underlying data source is used for the affected views and is it an extract or a live connection? 
 
Does the problem appear for all users or only for some?
 
Is the problem connected to one specific browser or does it appear on every browser (including Internet Explorer, Chrome, Firefox)?
 
When was the issue first noticed (since a recent update, 3 months ago, worsened over time) and what changes were made before the problem appeared?
 
Does the problem appear intermittently or all the time with the same workbook?
 
What is the approximate size of the Data being used for the problematic workbooks?
 
Does the performance issue also happen in Tableau Desktop or only in Tableau Server?

Please describe the exact steps that lead to the problem in as much detail as possible (if possible using screenshots).

How long does it take for the views to load (in seconds) and what is the desired loading time?

I understand that a live connection to BW HANA is used for the dashboard. Can you please create and extract of this data source and replace the live connection with it. We want to see if the live data source connection might be the culprit and by excluding all live connections we can confirm this if the problem persists with only extracts.