IBM Support

Performance improvement while using large sets of data from Netcool/Impact to populate DASH widgets

Question & Answer


Question

How to improve overall performance while using large sets of data from Netcool/Impact to populate DASH widgets?

Answer

If there are multiple widgets on the same page, all of them being populated with data sets or policy results from Impact, you might experience some slowness in loading that page.

In this situation you should check the following to try to improve the performance:
1.  Maximum number of widgets per page supported by DASH is 8-10. If you have more, you might consider redesigning your dashboard pages.
DASH supports event wiring, therefore, you can try to reduce the numbers of widgets by using this mechanism. You can design a scenario in which you can select a row in one table and send this selection event to a second widget (such as a SmartText widget or any other widget) and configure it to retrieve the complete alarm information details through another Impact policy.
For additional information you can check the following resource:
https://www.ibm.com/developerworks/community/blogs/69ec672c-dd6b-443d-add8-bb9a9a490eba/entry/JazzSM_Dashboard_Widgets_Interaction_Eventing_Demystified?lang=en
2.  On Impact side, check impactserver.log for any failed SQL queries, perhaps there are SQL queries that are not valid or have not been configured correctly and they are generating different error message. If this is the case, fixing such SQL queries will most likely lead to improve the overall performance.
3.   If you use tables with a lot of column, you can also try to tune the data type to better serve the widgets by:
-  removing unnecessary columns from the data type table definition because this will reduce the size of the SQL query and results
-  applying a Data filter to proactively filter out rows you don't actually need
You can also consider breaking up the table into a set of tables (you can split them by categories, for example). This would improve usability and help distribute the load across multiple widgets.
4. If your trying to scale up the row count, then you must use the large table model feature from Impact.
For additional details about the required configuration, you can check the following resource:
https://www.ibm.com/support/knowledgecenter/SSSHYH_7.1.0.16/com.ibm.netcoolimpact.doc/solution/uidataprovider_lrg_datamodels_c.html
There are two ways to enable the large model support:
4.1. If the alarms are being retrieved from a single table, you can create a data type under your data source and tick the "Access the data  through UI data provider" checkbox. This will automatically enable access via the UI data provider and the large model support immediately but it is limited to a single table and won't allow you to modify the SQL query or use a policy to edit the data.
4.2. Implement a large model policy. For additional details and examples you can check the following resource:
https://www.ibm.com/support/knowledgecenter/SSSHYH_7.1.0.16/com.ibm.netcoolimpact.doc/solution/imsg_largedatamodelsupport_uidataprovider_policies.html
This involves changing the policy to handle all the paging, filter and ordering logic.
Please note that the large model support only applies to table/list type widgets. Charting widgets don't currently benefit from the paging model offered by large model support.
 
5. It is also recommended to avoid turning up the debug levels as logging excessive details of the UI data provider can have a minor impact on overall performance as well.
6. Check you Impact and DASH product versions. With new fix pack being released additional bug fixes and enhancement are being introduced.

For Netcool/Impact in particular there were several APARs fixed starting with Fix Pack 16. One example is the below one:
IJ14490: WHEN DASH WIDGET CONSUME UI DATA PROVIDER OF DATA TYPE, IT KEEP QUERYING EXTERNAL DATA SOURCE:
introduced in Fix Pack 16:
https://www-01.ibm.com/support/docview.wss?uid=swg1IJ14490

[{"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSSHYH","label":"Tivoli Netcool\/Impact"},"Component":"","Platform":[{"code":"PF016","label":"Linux"}],"Version":"7.1.x","Edition":"","Line of Business":{"code":"LOB45","label":"Automation"}},{"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSEKCU","label":"Jazz for Service Management"},"Component":"","Platform":[{"code":"PF016","label":"Linux"}],"Version":"1.1.3.x","Edition":"","Line of Business":{"code":"LOB45","label":"Automation"}}]

Document Information

Modified date:
30 September 2019

UID

ibm11076979