EWA section 7.2.1 DB Load Profile

[Former Member]

Hi,

I have a question regarding DB Load profile in EWA.

The section number might be different for you.

It starts saying "The following diagram shows the DB load caused by dialog, RFC, HTTP(S), and background tasks, over different time frames."

"The data provided in the diagram represents the average number of database processes occupied by each task type in the database during the specified time frames."

but then the graph is saying something completely different.

Average No of Work Processes Waiting on DB response.

so what does that mean? a shortage in resource during the night? or high utilization of RFC Wps that must be compared to total number of RFC Wps available?

Thanks for any input.

Regards,

Adel

[Former Member · Accepted solution]

Hi Adel,

Looking at the graph my first suspicion would be slow response time from the DB or blocking transactions.

I would start by looking with ST03, ST04, ST06 and generating a ASH/AWR report for the time period in question to see what's happening on your DB/OS during the time window when you have a high number of RFC waiting on DB response.

In my experience waits on DB more often than not come down to:

Resources: (IO,CPU,MEM)

Poor Selects

Indexes: Poor choice, poor design, missing indexes.

Locking (exclusive locks)

Over zealous parallelism.

KR,

Amerjit

[Former Member · Accepted solution]

Thanks Yuksel.

To me there is  a difference. When I read DB Load and hours, I am expecting to see that between 00 and 01 it was mainly RFC then during the working hours mainly DIA etc.

Like some statistical data but now I understand that it is simply long running task and not really DB Load.

My question was not how to analyse it but more about the information provided itself.

At first I thought we had a lot of RFC and maybe shortage.

After checking with Wily we are fine and we do not need to change OP Mode.

This graph was confusing me. but as highlighted by Amerjit, it is more about long running task.

So yes first ST03 to see what is running and maybe STAD as well.

Thanks Both.

Regards,

Adel

[yakcinar]

Hello Adel,

The graph shows # of WPs that are waiting for DB process to finish the DB requests.

There is no conflict betweeen graph and header in my opinion.

And there are detailed explanations too.

At night time (bw 2 and 6) there are more RFC works in your system so keep # of DIA wps more than BGD wps. (As you know normally BGD wp numbers are increased during night times) You can use operation modes and dynamic wp distribution functionality.

Investigate # of active RFC wps between 2-6 using /N/SDF/MON transaction.

If they are not enough regulate your wp counts using operation mode accordingly.

Regards,

Yuksel AKCINAR