Performance Issue while SAP_ALERT_CALCULATION_ENGINE

[Private_Member_19084]

Hi everyone,

we are using Technical Monitoring on SAP Solman 7.1.

And now we are facing a big performanc eissue (HTTP and also Dialog).

Days before we didn't have this problem.

But I did not change anything. I did only remove the monitoring of a few managed systems and assigned exactly the same again.

But the same checks are active now.

I have the feeling, that the performance issue always happens when the job  SAP_ALERT_CALCULATION_ENGINE is running.

But not more than 4 jobs are running.

Can someone guide me how to analyze the problem?

I did also already check SLG1 of the report, however the amount of checks are similar than in the past - only the time did increase.

Kind regards

[Private_Member_19084 · Accepted solution]

Problem seems to be solved now.

We have changed parameters:

zcsa/table_buffer_area to 320000000 (old 150000000)

rtbb/buffer_length to 120000 (old 60000)

And I've also reduced parameters for extractor-parallelism in Solman_workcenter->Infrastructure->Extractor Configuration.

[Private_Member_19084]

Hello again,

so the performance issue is still not solved.

Now I did deactivate all the technical monitoring.

Did shutdown Host Agent on Solman.

But still the system is very slow (dialog and http (servicedesk)).

But we don't have any memory problem and I don't know how to analyze the problem further.

Can someone please help me analyzing this problem?

Thank you

[Private_Member_19084]

Hello everyone,

I've now found out the real problem.
It seems to be the job "EFWK RESOURCE MANAGER" (report E2E_EFWK_RESOURCE_MGR).

Always when I execute this job I get the problem.

I've now also reduced the numer of used dialog processes for extractor in work center (note  1847293).

However, the performance is still very bad.

OSS-message is opened. But maybe one of you knows a solution also.

Thank you

[Private_Member_19084]

OK, this did not solve the problem.

What do you think, may it be the problem,that  I did deactivate the BI-Forwarding of some metrics?
But I thought, than it should be faster.

Kind regards

[Private_Member_19084]

so, in DBACOCKPIT  I can see that a lot of locks are held are done (around 1000 per minute).

May this cause the problem?

[roland_hennessy]

Hi Christian,

Can you see if this latest SAP Note 1964945 helps here?

Kidn regards,

Roland

[Private_Member_19084]

Hello ROland, I think thats it.

Thank you for the note.

After implementing the response-time did fall down.

Kind regards

[Former Member]

Hi,

we faced these issues as soon as we activated some monitoring options. but we managed by following the sap note 1835721 - Troubleshooting performance problems in Solution Manager

you can check the other note 1928160 - MAI: Performance der Ereignisberechnung optimieren or 1869134 - Performance improvements for MAI repository SP8

though you have not more jobs, if the note 1861580 applicable in your system, implement the note and check.

Thanks

Jansi

[Former Member]

Hi,

Refer this document for Sybase on setting up Automatic Table Maintenance in Sybase. This will take care of updating the statistics of all the tables.

http://scn.sap.com/docs/DOC-15162

Alternatively you may follow the steps below to update statistics of individual tables/index.

To check missing stats  execut isql -Usapsa -S<SID> -X -w999 -i checkstats.TXT -o checkstats_ot.txt
    This will generate list of table having missing statistics. Information will be recorded into the text file checkstats_ot.txt

Thanks

Rishi Abrol

[Private_Member_19084]

I've now implemented note 1869134 and 1717403.

Hope one of them solves my problem.

[Former Member]

Hi,

What is the DB. Have you tried to pull the db stats of all the tables and see if it helps.

What is the SP level of the solution manager...

Have you check the below note..

1712812 - Performance of job SAP_ALERT_CALCULATION_ENGINE

1861580 - More then four parallel jobs SAP_ALERT_CALCULATION_ENGINE

But any how you are not running the jobs more than 4.

1932792 - Error "aggregate for managed object not found" with Solution Manager 7.10 System Monitorin...

Thanks

Rishi Abrol

[Private_Member_19084]

Hello Rishi,

we are using DB Sybase.

Solman 7.1 SP 8.

I've checked all 3 notes and no one is valid for us and this case.

What do you mean with "But any how you are not running the jobs more than 4".

Does it mean I should do this?

Kind regards

[Private_Member_19084]

One important further info.

I did deactivate the "BI-Functionality" in Data-Collection of some metrics.

Seems them I have a lot of entries in table MES_DB_AGGREGATE (around 15 000).

I've already checked note 1785201 - but was not implementable.

Kind regards