Please share best practices about optimizing responsiveness for heavy Visual Analytics Reports

[acordes]

We are experiencing a long waiting time in sophisticated reports in terms of parameter usage, linked objects, big tables, aggregated measures and tables and filters. 

The admin team is working on implementing solutions and pushing towards upgrading to the latest release.

But I need SAS experts to share best practices.

I imagine and somehow have experienced the following do's and don'ts. But it is more gut feeling than real evidence:

Good practices and common sense:

• Apply data source filter when possible (sometimes it's disabled, under what circumstances?) and use a parameter for the filter condition. 
• Avoid VA joined tables if it could be done in the ETL SAS code section
• Do not carry unnecessary columns in the report's tables
• Use varchar for the tables' columns if possible

But I do not know if using a slider for a timestamp column with 100k distinct values on the top of the page slows down the report. 

Or which calculated items consume a lot of compute, like string operations. 

Or ¡f linked actions used throughout the report can cause latency. 

Thanks in advance. 

[Sam_SAS]

Hello,

The items you included are all good practice. I believe that in the past, high-cardinality prompts were a performance issue and we made a fix. I am not sure how much they affect performance today. I am also not sure about the specific case of date sliders, I think they might be less of an issue than a discrete control like a drop-down list.

We should probably add some tips to the documentation, so I will watch this thread.

Sam