Earn a 50% discount on the DP-600 certification exam by completing the Fabric 30 Days to Learn It challenge.
We have a Power BI Embedded dedicated capacity and a workspace enabled for the dedicated capacity.
Does this mean that push datasets never use our Power BI Embedded dedicated capacity? If so, why is that?
Hi @xli629b ,
Based on my research, we think it may use little resource that less than dataset which use Import Mode, because With a push dataset, data is pushed into the Power BI service. When the dataset is created, the Power BI service automatically creates a new database in the service to store the data. The push dataset only need to receive the new data comes in and then trigger a tile refresh every time new data is received.
But for Import dataset, it need to import entire dataset inside memory, then receive all the data and execuse some M code which may use more resources.
Please refer to those documents about push dataset:
https://docs.microsoft.com/en-us/power-bi/service-real-time-streaming#pushing-data-to-datasets
https://docs.microsoft.com/en-us/power-bi/refresh-data#understanding-data-refresh
https://docs.microsoft.com/en-us/power-bi/developer/api-rest-api-limitations
Best regards,
Thanks for the research. However, how does report filtering work? Does Power BI Embedded VM get the data from storage and then filter it in memory? Or is it handled by the database itself, which is on a machine other than our Power BI Embedded capacity VM? Does this mean that if we used Push Datasets we would need only a comparatively small VM to handle comparatively large loads?
Hi @xli629b - Based on a recent support interaction with the product group, I've been informed that Push Datasets use Shared Capacity, regardless of being created in workspaces assigned Premium Capacity. I was also informed there is no plan to change this behavior at this time.
This may explain why you see no impact on your capacity when accessing push datasets.
A side effect of this decision is that datasets in premium capacity workspaces with connections to streaming datasets within that same workspace will display an error when creating / displaying visuals or loading in Power BI Desktop:
The Product group suggested 2 workarounds: Creating the new dataset in a Shared Capacity Workspace (which would prevent App access to non-pro users), OR creating an import dataset based on the push dataset and refresh at regular intervals. This is not optimal either, but may be our only choice if we need to allow non-pro users to consume the data in an app.
I've requested that the documentation be updated to state the implications of the design decision, we'll see how that goes!
Good luck,
Joel