Skip to Content
Technical Articles
Author's profile photo Jisi Chen

How to Isolate a Performance Issue in SAC Stories?

Introduction:

In this blog, I will be giving a review about how we isolate a performance issue by following SAP’s best practice in combination of our experience in SAP Analytics Cloud Stories.


 Context:

There are many factors in SAP Analytics Cloud (SAC)  that might cause performance issues, where should we start and how should we isolate the problem?

  • Browser and Network

Based on our experience, the recommended browser of SAP Analytics Cloud is Google Chrome latest version. If your screen is running slow in Internet Explorer, New Microsoft Edge, Safari, etc. please make sure to use the latest version or switch to Google Chrome for best experience. Test the  screen with Incognito Mode. You may also want to use task manager to check the local memory and CPU usage. Switching to a different network might help identify the issue caused by local network. 

  • Backend Connections

HANA – If your story is connecting to a HANA model, please work with HANA developers to optimize the calculation view.

    1. Use aggregation before calculation
    2. Filter data as early as possible and hide unused dimensions/measures in the view.
    3. Run performance analysis mode for the calculation view in HANA

BW Queries – If your story is connecting to a BW query, please make sure your BW query is efficient 

    1. Reduce unused dimensions and hide key figures can be shown
    2. Within structures, filter starts with the highest level first
    3. Within structures, use user exit variables to calculate things like QTD, YTD
  • Inside SAP Analytics Cloud

In Models:

    1. Avoid Exception Aggregation in the model if necessary
    2. Depending on the use case, our recommended approach is to consider creating powerful calculations in the backend first then in SAC model.

Performance-Wise: Complex calculations in the backend > calculations in the model > calculations in the story

In Stories:

    1. Check if you are using canvas instead of responsive pages. Responsive pages are better especially when you have the future need to view story content in mobile.
    2. Check if you have any large size of image in the story, if so, please consider the following: SVG > PNG > JPG image file
    3. Be mindful of using linked dimensions in SAC to create blended charts or display data from multiple models. Blending models in SAP Analytics Cloud is recommended when you have a small data set. If your plan is to blend big size data from two different sources, please consider joining data in your backend system.
    4. If your story is connecting to HANA, please make sure to enable Query Batching under Query Settings on Edit mode.Query%20Settings(Source: SAP Analytics Cloud Query Settings Window)
    5. If your story is connecting to BW, please make sure to enable Query Batching under Query Settings on Edit mode.(Source: SAP Analytics Cloud Query Settings Window)
  1.  6. Restricted Measure

Restricted measure is a smart choice when you need to create a measure that is limited to a certain dimension selection (Ex. Revenue for Category named Orange Juice). But please note that a restriction cannot push down to the query level like a filter. Therefore, based on a specific use case, please identify the need between applying restricted measures and story filters.

In SAC Content Folder:

In late 2020, SAP Analytics Cloud has rolled out a SAC Content folder for system and BI admins. This content folder includes one sample story and three models about performance.

SAC%20Content%20Folder

(Source: SAP Analytics Cloud )

Now system administrators can monitor content performance via the Statistics and Analysis story or build a customized story based on these models. 


Conclusion:

With this blog we have gone through the possible factors to isolate a performance issue. Of course, these are not the only causes but enough clues for you to start. If you are interested in learning more, please do not hesitate to reach out.

Assigned Tags

      Be the first to leave a comment
      You must be Logged on to comment or reply to a post.