Report In A1 Embedded Capacity referencing RLS Dataset in Pro Workspace? Will this cause Problems?

[TravelMan]

Hi All,

I've got a 500mb data set that expands out pretty large whilst refreshing that we've been using within the organisation with pro licence user model for a year or so now.

This sits in a pro workspace and refreshes a few times a day perfectly fine.

The aim is to embed a report sitting on this dataset into our customer portal using application user control.

I've spun up an A1 SKU Embedded resource and created workspace connected to that resource, I've published a report that references the dataset in the pro workspace and published the report to the Embedded A1 Workspace which works fine. The RLS identities requests and embedded token fetching all work fine.  So all looks well on the face of things.

My main question is are we going to run into unforseen problems downline with this configuration?

Hypoethetically expecting a max of 20-30 simultaneous users at peak times.

I tested moving a mirror of the dataset to the Embedded workspace itself and refreshing, but the refresh immediately caused an out of memory error when sitting in the A1 SKU.  So obviously that dataset can't live within the embedded workspace. A2 is within budget but anything above that means looking at other options. The refresh will also max out an A2. Preference is for the dataset to stay where it is in the pro workspace with the reports in the embeded workspace.

My questions are :

Are we going to experience any issues by having the report in embedded capacity referencing a dataset in a pro workspace?

Because the embed request touches a report in the embedded workspace but a dataset in the pro workspace is this    going cause a limitation on the number of embed tokens that can be generated because there is a non embedded                           entity involved? Or because the report itself is in the embedded capacity is that the primary concern of the embedding token limit?

Regards,

Chris

[Anonymous]

HI @TravelMan,

AFAIK, power bi premium is capacity based on license. This content stored in the shared capacity only has pro features.

Types of licenses for Power BI business users - Power BI | Microsoft Docs
If your embedded contents are not hosted in the premium capacity, this means the premium feature (view content without a pro license, refresh amount...) does not work.
Also, the embed token for premium is not limited but the pro license only includes a few trial tokens for usage. So your users may be faced with token exceed to the limit.

Regards,

Xiaoxin Sheng

[TravelMan]

Hi Xiaoxin, Thanks for the response.

I think i've worked out how to test my theories and structures.

The GetAvailableFeatures by name API with the Keyword "embedTrial" seems to be a good indicator of what configs trigger the Trial tokens to be depleted. 

https://docs.microsoft.com/en-us/rest/api/power-bi/available-features/get-available-feature-by-name

The annoying thing is the dataset itself would happily work in an A1 / A2 capacity server. Its only the refresh process which causes the problem. Which is odd as I would expect a paid capacity to have better refresh limits and capabilities than a pro workspace.

Is there anyway to get an A2 capacity to refresh a large dataset that exceeds the RAM limit.

It seems odd for an azure premium capacity SKU to be so limited on the refresh stage.

A manual refresh on Desktop an republish would work, but its not a process i want to create.

[Anonymous]

Hi @TravelMan,

In fact, A2 did not increase the refresh amount. (Pro 8/ day, Premium 48/ day) If you are limited by pro reefer and out and you can't move all content
For Premium capacity, you can try to host them online or cloudy, then power bi will use the different refresh methods and not have refresh limitations on this type of data source. 

Data refresh in Power BI - Power BI | Microsoft Docs

Regards,

Xiaoxin Sheng