Azure Analysis Services Model Cache

[SHHargrove]

We are configuring PowerBI to use an Azure Analysis Services data model.  Looking at the different service levels for Azure Analysis Services, they are expressed in terms of QPU's and Model Cache.  What does "Model Cache" mean in this context?

Is it the total size of all the models on a server, without regard to whether any or all of them are active at any point in time?

If so, that's predicatble.

Or is it some other measure like the total memory consumed by analysis services to service the model and any requests?  If it's something like this, I'm not sure how we can properly size our Azure Analysis Services server.

[v-yulgu-msft]

Hi @SHHargrove,

According to your description, you want to load data from Azure Analysis Services into Power BI, right? Now you want to confirm how to configure your Azure Analysis Services server correctly so that it can hold data model with great performance, right?

As your questions are more related to Azure Analysis Services, I would suggest you post these questions on Azure Support Community so that you can get more profession advice.

Here are some basic knowledge anout connect Power BI to Azure Analysis Services:

Connecting Power BI to an Azure Analysis Services server

Creating your first data model in Azure Analysis Services

Best regards,
Yuliana Gu