Roadmap to Analyze Performance Problems in an SAP system

Spread the love

Performance of the SAP system depends on many factors. Whatever may be the reason for performance bottleneck is, the high dialog response time is the indication of the problem. So by analyzing dialog response time, we may find out the solution for the problem quickly. We should have detailed knowledge on dialog response time components to analyze. We have to consider Following points while analyzing dialog response time to find out the potential performance problem in an SAP system.

  • High wait time occurs, when there are no available free work processes after receiving a request from the dispatcher. So sufficient no of work processes should be configured to handle the requests more quickly.
  • ¬†High roll-wait time occurs when there is a communication problem with GUI, external system or when large amount of data requested.
  • High database request time occurs when there is a CPU or memory bottleneck on the database server, communication problem with the database server or because of expensive SQL statements, missing indexes, database locks, missing statistics or small database buffers configured.
  • High CPU time occurs when there is an expensive ABAP processing, such as processing large tables and inefficient programming.
  • Processing time may be more than the twice the CPU time occurs when there are any CPU bottlenecks

By analyzing and rectifying above problems, we can improve performance of the SAP system. There are many important views in the workload monitor (ST03N) and then Analysis views to analyze the statistical data to find out the exact performance problem. I will explain more details about workload monitor in the next article.

(Visited 662 times, 1 visits today)

1 Comment

  1. Oscar

    Some time ago we suffered from problems in performance in our SAP system. We tried to increase the work processes available, but it became a very bad solution. We noticed that after increasing the work processes led to bottlenecks in database. Finally, the solution was buying a physical disk.


Leave a Reply

Your email address will not be published. Required fields are marked *