Azure Analysis Services - refresh

New Contributor


our current setup is on-premises PostgreSQL DB that serves as data warehouse. We use Power BI Pro platform to create data model and refresh the data daily, using locally run Data Gateway. As the size of the data is already resulting in refresh times longer than 1 hour, and getting close to the 1GB data model size limit, I am considering setting up Azure Analysis Services.


However, what is not clear to me is how the data refresh is handled in such case. I presume that on-premises data gateway would still be required in order to access our PostgreSQL DB. How is the refresh (and in addition, incremental refresh) handled when using Azure Analysis Services? Is there a scheduler available? Are there some limitations (e.g. how long a refresh can take before it timeouts?)

Thank you.

1 Reply

Hello _Vladimir_,
Azure Analysis Services uses the same data gateway as Power BI. Here are the docs 
Incremental refresh would be defined in your partitioning strategy in how you set up your model. 

The refresh would be handled outside of Analysis Services via your existing ELT process and an XMLA command, or an Azure Data Factory pipeline or Azure Runbook


If you use just regular processing in Azure Analysis Services the processing may still take longer than an hour due to the long running HTTP call. You can get around this in AAS by using asynchronous refresh instead.