Solved: Re: Dataset/Data storage capacity

Anonymous · ‎07-10-2019

In my scenario I have configured a Stream Analytics to write in a pbi dataset (PRO license).

A dataset is automatically created with the historic data analysis option enabled.

I've found many informations about capabilities but not always congruent with each other.

In this documentation page it's specified 5.000.000 max rows stored in a non retention policy dataset (it seems I cannot configure it using Stream Analytics).

In this page it seems the maximum number of rows depends on DirectQuery (I can image we are not using DirectQuery because of Stream Analytics).

this page says users with a Power BI Pro license can create app workspaces, with a maximum 10 GB of data storage each and there is a 1 GB limit, per dataset, that is imported into Power BI.

Which is the right interpretation of these information? Does the dataset store new data using FIFO retention policy?

Are these considerations applied to CosmosDB data source too?

Thank you

Anonymous · ‎07-10-2019

This is what applies to your scenario:

In this documentation page it's specified 5.000.000 max rows stored in a non retention policy dataset (it seems I cannot configure it using Stream Analytics).

Does the dataset store new data using FIFO retention policy? Yes

Are these considerations applied to CosmosDB data source too? To streaming datasets, Cosmos falls into the 10GB as it is not a Direct Query supported data source.

For the streaming if you want to save the data you need to split it off and have stream analytics dump the data in a storage location in order to retain it. The cool thing with this scenario is that you can use Power BI to recieved the live data and show it in a dashboard, have a report that uses the FIFO of the retained storage, then connect other reports on the storage location where you would be piping the SA data to.

View solution in original post

Anonymous · ‎07-10-2019

This is what applies to your scenario:

In this documentation page it's specified 5.000.000 max rows stored in a non retention policy dataset (it seems I cannot configure it using Stream Analytics).

Does the dataset store new data using FIFO retention policy? Yes

Are these considerations applied to CosmosDB data source too? To streaming datasets, Cosmos falls into the 10GB as it is not a Direct Query supported data source.

For the streaming if you want to save the data you need to split it off and have stream analytics dump the data in a storage location in order to retain it. The cool thing with this scenario is that you can use Power BI to recieved the live data and show it in a dashboard, have a report that uses the FIFO of the retained storage, then connect other reports on the storage location where you would be piping the SA data to.