When to use Analysis Services in Power BI?

[schilders]

Our Power BI environment utilizes On Premise Gateways to refresh semantic models.  We've clustered the gateways and load balanced them.  A new dashboard development project returns in excess of 30 million rows (medical claim detail line items) as the primary object in a data model that implements the star schema.

I'm curious to know if we have reached the point where SSAS multidimensional cube might be beneficial for this type of data?  Can anyone provide insight as to why we would not want to implement this new direction in our dashboard development?  Are there any pitfalls to evaluate when considering use of an OLAP Cube instead of related business objects that contain their own embedded Transact SQL queries?  We run SQL Server 2019 in a cloud-based environment.

Thanks in advance for any / all words of wisdom.

[lbendlin · Accepted solution]

and load balanced them

err, don't do that. The Power BI service will do that already. Don't mess with it.

Please note that Power BI is based on SSAS Tabular, which is a refinement/departure from OLAP.  You would take a step back if you chose OLAP over Power BI, including the substantial overhead of having to translate between DAX and MDX back and forth.