DX Unified Infrastructure Management

Hi Experts,

We are having UIM 8.1 setup. We are integrated our ITSM with Nimsoft alarm API and using USM alarm console. We are feeling slowness  and delay on updating alarm in USM alarm console. This is impacting our logic on ITSM. Engineer couldn't resolve the ticket even though alarm cleared. How to improve performance on USM alarm_console. Could anyone describe how USM alarm console is working?. 

Thanks and Regards

Ponnuchelvam.V

HI, I believe this would depend on the version you are running. the deprecated alarm console in 8.31 and earlier used to make calls to get the information from the nas sql lite databases. Usually with the *.db files in the nas directory got to be too large this would impact performance. This was usually the case with clients with a large number of alarms with high counts.

If you are talking about newer versions of Alarm console say in 8.47 this is pulling from the back end NAS_alarms table.

This is updated by the nas probe in the nas_bridge part of the code.

usually the slowness here again is seen when there is a large number of alarms with high counts.

Once way to try and solve this short term is to deactivate nas and alarm_enrichment probes and drop the nas_ tables from the back end and then activate alarm_enrichment and nas.

This will cause the tables to be recreated and the data from the local nas to be sync back up.

you could also look at the indexes and fragmentation on the nas_ tables and the cm_ tables and make sure these are less than 30%.  The underlying issue could be simple database performance.

Hope this helps

Hi Gene,

Thanks for your response. i agree that we have huge data and alarm processing per minute in my environment. Here, I am talking about USM alarm console and not about deprecated alarm console. We could see delay in alarm update only in USM alarm console and not in deprecated alarm console. I heard from CA support USM alarm console fetching alarms from UIM DB(backend) not from sql lite DB. We are doing indexing every week for all tables.

Do you have any idea How to improve USM alarm console performance other than indexing?.

Is there possible to change alarm api interaction to sqllite(nas) instead of refer backend(UIM DB)?

Thanks and Regards

Ponnuchelvam. V

there is no way to change the patth pf the API as to where it is pulling information.

If you have a high alarm and data volume indexing once a week is not enough it should be done daily

Might want to check the data in the nas_transactionlog and nas_transactionlog_summary and make sure this

is having data removed on a daily basis.

If it is not this would be caused by poor indexes and the the reason you updates are slow.

To correct this I would deactivate nas and alarm-enrichment

drop the nas_ tables on the back end database

activate alarm_enrichment and nas and then make sure the indexing is being done nightly.

If you are not on an enterprise version of the database that can do reindexing online you may need to look at that change as well.