Skip to Content
avatar image
Former Member

Response time issue..

Hi Gurus,

We are facing a problem with Response times with our APO Server. Avg Response time( for Dialog task type) of 1.2sec or below is acceptable . But we get response times twice the accepable limit..somtimes goes upto 3 sec's... When I check the expert mode -> for day then -> Transaction Profile -> Standard and sort it with Avg Response time for Task Type - Dialog I get the following result...

Report/Transactionsteps T Response Time Avg Response time

/SAPAPO/SDP94 218 1,073 4,923.8

SE16 41 201 4,890.9

ZCCR 36 152 4,235.9

SESSION_MANAGER 242 1,024 4,233.1

Note: All Times are in Milli Seconds.

My question is how can the above transactions have such a high response time. And how can I get to the crux of the problem..



PS: useful answers will be rewarded.

Edited by: Praveen Shankam on Mar 5, 2008 2:52 PM

Add comment
10|10000 characters needed characters exceeded

  • Get RSS Feed

2 Answers

  • avatar image
    Former Member
    Mar 05, 2008 at 04:31 PM

    To find out the exact reason for your performance problems you will need to run an ST05 SQL trace, run the transaction and see where the bottleneck is

    As for your individual ones:

    • SE16 is a database anaysis tool and 4.8 seconds is very good
    • ZCCR is a custom transaction thus you will need to trace (ST05)
    • I have never seen SDP94 so I cannot comment
    • SESSION_MANAGER is normally the sending of graphics when you first log into the system.  there is an OSS Note on SESSION_MANAGER as well, but in general, you can ignore this one.

    Russell Hull

    Message was edited by: Sylvia Delcheva

    Add comment
    10|10000 characters needed characters exceeded

  • avatar image
    Former Member
    Mar 07, 2008 at 03:24 PM


    Identify the Z programs first and do a SQL analysis. Run SE30 and know, how much time is spent on DB and how much time on ABAP. You would need what are the variants you need to input to run SE 30.

    Then run a SQL plan, and see the bottlenecks on where exactly the problem is happening. Like, it may be an issue with Expensive queries. Abap program may read a single entry from millions of records.So, identify the exact place of expensive query in the program and forward it to ABAPers.

    In your case the system response is moderate. So, try to focus on other areas like parallel users logging in, and try to distribute them accross app servers using smlg.

    And also, dont start any measure with three or four days of data, have atleast four months of past data, review the most expensive parts and then start analysing.

    Hope it helps, please let me know if you need anything more.



    Add comment
    10|10000 characters needed characters exceeded