Leveraging Usage Analytics for Determining Heavy Widgets
In this article, we can find a suggested approach on how to determine dashboard widgets that consume a lot of memory and may cause performance issues.
Usage Analytics
Usage Analytics is part of this solution; extending the default number of days stored (the default is to keep only 30 days of usage) is optional but will provide a clearer status of your environment.
1. Enable Sisense Usage Analytics - Follow this article to enable Usage Analytics.
https://docs.sisense.com/main/SisenseLinux/enabling-usage-analytics.htm#
2. Optional : Extending the number of days stored (Retention [Days] property) as described in the article above,
* Warning ! Increasing the number of days to a higher number than 30 days might affect server performance. Mainly during the usage Analytics cube build. It is strongly recommended to test the cube's stability, capacity, and performance before scheduling it for a frequent build.
3. Open 'Usage Analytics' > 'Usage - Dashboards' dashboard
4. Locate and duplicate the 'Dashboard Usage' widget
5. Edit the widget and add the following Rows:
- widgetName
- widgetId
- queryResponseLengthBytes
6. Rename the rows if required
7. Sort the widget by queryResponseLengthBytes in descending order
As a result, you will get a list of widgets that consumes most memory per dashboard and process them accordingly (re-design, add filters to limit data, etc)
Here we found how to enable the Usage Analytics feature and modify a pre-defined dashboard to get a list of widgets with memory usage details