Skip to Content
Technical Articles

SAP Analytics Cloud Performance Analysis for SAP BW Live Connection

Analyzing the performance of SAP Analytics Cloud stories requires the analysis of different components: SAP Analytics Cloud story performance, possible network delays and of course the SAP BW queries in case of a live connection.

SAP KBA 2511489 gives a great overview of performance related topics and of course you should follow the best practices for performance.

But sometimes a constellation with a specific story and the used BW queries needs to be analyzed and you would like to analyze the BW performance on a granular level as in the RSRT query monitor. That’s what this blog post is about: analyze query runtime in SAP BW in the following three steps.

1. Record story execution requests

First, we need to record a trace of the requests sent from SAP Analytics Cloud to SAP BW. This can be achieved quite easily using the Google Chrome or Edge Chromium developer tools. The procedure is described in SAP KBA 2280022 . The result is a HTTP archive (HAR) file that contains all the requests sent to the InA interface in SAP BW.
You will also find an approach using Fiddler, but from my experience using the developer tools of the browser worked just fine.

2. Import and trace requests

To analyze the trace, we now log into SAP BW and open the InA Test Monitor (transaction RSBITT):

The InA Test Monitor can be used to analyze and execute requests sent to the InA interface and create RSTT traces for further analysis. Both with your own user and with other users, just like the “Execute BW query as” function found in transaction RSECADMIN/RSUDO.
This can be selected in the options by unchecking the options “Use SSO” and “Current user”.

Now the HAR file needs to be imported into the monitor:

The monitor will now analyze the contents of the HAR file in a tree structure to allow the analysis of the requests:

Here you can also see all the filters/selections that are sent to SAP BW. But I wanted to show how to analyze the performance and for this we need to activate tracing:

Now we are all set to execute the requests and we can do this either step by step or by running the whole sequence of requests:

I suggest running the whole sequence because with this you can already “feel” how long it takes SAP BW to process the requests. Once the requests have been processed, the Trace ID is available, and we can go on to step 3:

3. Analyze BW query execution

The last step is to go to the RS Trace Tool (transaction RSTT) and analyze the query execution with the known options like “Display Statistics Data”:

And from here on the known analysis and optimization of BW query performance starts! For this  SAP KBA 2455142 and SAP note 1681396 provide further information.

Conclusion

Using the InA Test Monitor allows you to analyze the response time of the requests that are sent to SAP BW from SAP Analytics Cloud and identify the reasons for performance issues. Unfortunately, it does not (yet?) provide an automated solution 😉

2 Comments
You must be Logged on to comment or reply to a post.
  • Thanks for this, Jens! We are actively troubleshooting poor story performance but from the SAC mobile app perspective. Is there a way to capture something like the HAR from the Analytics app for performance analysis?

    Note that you have a critical typo in the transaction code. The InA Test Monitor transaction should be RSBITT (the RS prefix is common in BW representing “reporting server” as legend has it)

    This wiki also seems helpful for RSBITT: https://wiki.scn.sap.com/wiki/display/BI/Transaction+RSBITT+-+InA+Test+Monitor

    Thanks!

    Doug

    • Thanks for the hint Douglas – I have corrected it!

      Unfortunately I am not aware of any option to record a trace with the mobile app. But measuring the response times for mobile consumption will only differ in regards to to additional network or mobile rendering. To analyze the BW part of this it should be sufficient to reproduce the same navigation steps in the browser. But maybe I am seeing this too simple?