Understanding ABAP Traces

[former_member383774]

Hello everybody,

i'm trying to improve the performance of a prod system using ST12. It executes correctly and i get loads of data. The question is now, how to understand/interpret the data? It looks kind of obvious to sort the results and look for the tasks that take longer, but if none of them really stands up, what else can we do?

Thanks in advance,

Joaquin

[former_member383774]

Hello everybody,

i'm trying to improve the performance of a prod system using ST12. It executes correctly and i get loads of data. The question is now, how to understand/interpret the data? It looks kind of obvious to sort the results and look for the tasks that take longer, but if none of them really stands up, what else can we do?

Thanks in advance,

Joaquin

[Former Member]

Please check OSS Note: 755977 which talks at length about various attributes in ST12.

[Former Member]

please check my standard recommendation, there are 2 blogs from me on SE30 and ST05

SQL trace:

SE30

I know that you use the ST12, but the ST12 is nothing new, it is on top of the two traces and

nice tow operate them both, but results are the same.

Siegfried

[keiji_mishima]

Hi Joaquin

In addtion to ABAP Trace of ST12, did you also check STAD records ?

How is the time share of ABAP trace looks like ?

In general, if DB time is higher (let's say more than 60% of total response time), you need to check DB access path and ABAP DB access code from ST05 (SQL trace). If CPU or processing time in STAD or ABAP time in ST12 is higher (let's say more than 60% of total response time), there is a possibility of tuning potential on ABAP coding like READ statement/Loop at ITAB with WHERE condition / Delete where statement.

In you description, should mentioned, you could not see any particular time consuming part. But I little bit wonder did you take ST12 trace with internal table option ? This option become important when you have high time in ABAP time (or CPU/Processing time in STAD). Without this option, you cannot see time consuming READ/LOOP operation. So if you had higher time in ABAP time, and you did not take ST12 trace with internal table option, please take ST12 trace again with internal table option.

After that, please display ABAP trace by sorting NET time.

if you still do not see major part, how about checking time consuming module ?

Different from SE30, ST12 can provides ABAP trace per module and call hierarchy (even with aggregation). So if you find out time consuming module pool (in this case, please check gross time), please check the module pool and its content. <If the module is SAP standard code, please search node with time consuming module pool>.

another idea is, display call hierachy of most time consuming code by hitting where used icon of ST12. Then you can see call hierarchy. Sometime, due to too many call, some part of coding become very expensive. in this case, by using call hierarchy option, you can see how the program is called. And you may find out the reason why the program is called so many times (sometime, by doing this, you may find due to program bag, too many times, the program is called....etc).

Best Regards

keiji