Data Geek Challenge 2 – ST22 Dump Analysis using SAP Predictive Analysis

Programmer always wants his/her code to be dump free. However, sometimes during unit testing we get some dumps which can be viewed from transaction ST22. One day I was also doing testing and suddenly program dumped and then went to ST22 and analyzed the logs.

Suddenly some questions came to mind and that was How many dumps we get/produce J daily in a particular SAP landscape? Is there any specific dump we are getting for particular time frame? And many more…

To answer these questions, SAP Predictive Analysis can be used to decide what can be done to avoid those in future. To do so, we will need a dataset. I have used following dataset.

Dumps from ST22 from August 01, 2013 to August 19, 2013

As the data downloaded from SAP cannot be uploaded as it is, we need to format it. The file after formatting looks like as follows:

Later simply upload data using any of the below data source.

Define some attributes as measures. Drag measure on Y-Axis and then some other fields on X-axis and then select different views, graphs, pie-charts etc. and do your analysis.

Let’s see various scenarios which will answer above questions.

Scenario 1: Daily Runtime Errors in SAP System from Aug 1, 2013 to Aug 19, 2013

We can see that total number of runtime errors day wise for a particular time in below graph. We can easily found out on which day dumps were maximum.

However this is still not helping us in identifying by which user id we are getting maximum dumps on various days. Let’s see next scenario for this.

Scenario 2: Date and User ID wise Runtime Errors in SAP System 230 (Aug 1, 2013 to Aug 19, 2013)

From below screenshot we can easily find out that on August 1^st 2013 there were maximum dumps by user id SysuserALE and total errors were 570. Apart from that we can also find out date and user wise dumps.

Scenario 3: Maximum Runtime Errors by USER ID

Runtime Errors user wise in terms of Pie Chart

Runtime Errors user wise in terms of HEAT MAP

Runtime Errors user wise in terms of TAG CLOUD

Scenario 4: Maximum Runtime Errors in SAP Landscape 230 by runtime errors name (Aug 1, 2013 – Aug 19, 2013)

In scenario 3, we saw maximum errors by a particular USER id. Let’s see maximum errors by error name itself. “RFC_NO_AUTHORTITY” error was highest in number during mentioned time frame.

Scenario 5: Maximum Runtime Errors by Date

Now let’s find out maximum runtime errors date wise in RADAR Chart.

Till this point we were just analyzing errors by user id, by error wise, date wise and overall daily how many runtime errors we get in SAP Landscape.

In next scenario, we will see is there any trends in occurrence of these errors?

Scenario 6: Runtime Errors Trend in weekdays and weekends

As name itself suggest, we can see that runtime errors coming on weekends or holidays is significantly getting reduced while it’s getting increased on weekdays.

Another view for similar trend.

Till this point we were just doing the analysis of available data in various formats such as Graph, 3D views, Pie Charts, Geographical view etc. but SAP Predictive  Analysis also provide option to predict results based on data. Following are some steps for predicting results.

Click Yes. It will take you to Results tab which has 3 different tabs. Grid, Charts, Visualize. Since this is not good data for prediction, results are not appropriate. Please ignore the results. I have added these just for reference.

Closing remarks:

There are many other views/charts available in SAP Predictive Analysis tool such as geo-graphical view but that can be used for a dataset which involves measures such as Country, state etc.

If you want to see that feature, please refer my blog at SCN:

http://scn.sap.com/community/lumira/blog/2013/08/15/data-geek-challenge–district-wise-season-wise-crop-production-statistics-in-india-from-1998-to-2003-by-using-sap-predictive-analysis

Note: This is just one simple example and does not demonstrate all the functionality available in SAP Predictive Analysis tool.

I hope you like this blog and look forward for your comments and suggestions.

Thanks for reading!!!! 🙂