An error occurred in Power BI Premium backend services

[odumna]

Hi
I got error "An error occurred in Power BI Premium backend services" on report that had worked fine till yesterday (or fome days ago)

Activity ID: 0a8089ca-8ee5-4cf6-96ea-37481f63adbc
Request ID: d29f753f-c0fa-4dfd-9f93-9d4238788577
Correlation ID: d106e310-a32d-8f79-1292-08151a1a0ca1
Time: Fri Feb 10 2023 11:12:05 GMT+0200 (за східноєвропейським стандартним часом)
Service version: 13.0.20006.56
Client version: 2301.5.12183-train
Cluster URI: https://wabi-west-europe-redirect.analysis.windows.net/

I created and published report several months ago, and it worked fine.
But yesterday users notified me about issue "An error occurred in Power BI Premium backend services". 
On Desktop this report successfully refreshed and shows data, however on portal I get error.
I tried to investigate the reason and found that problem is in measure that based on direct query from other dataset. Once I delete this measure from visual all works fine.
But I need this measure and I need it from direct query of another dataset.

Could you please fix this issue on your side

[Kellef]

I set up a gatewayconnection to my masterdataset as cloud connection from my subdataset, i got it directly and when i removed it it dissapered. I dont really how to set up a scheduled refresh now without getting this problem again, anyone got a soloution for this case?

[Amelie]

Hello, 

I got the same error with a composite model. The solution for me was to not use the "bind to parameter" and not join a parameter to a column of a table. 

[supriya1909]

Hi, Did you get any resolution for this?

[gnico]

Hi,

I have been encountering this issue ("An error occurred in Power BI Premium backend services") today on a report.

The dataset tied to this report is a composite model and has a Direct Query connection to a remote dataset.

So I have a local and a remote dataset (according to Microsoft semantics in the following documentation: https://learn.microsoft.com/en-us/power-bi/transform-model/desktop-composite-models)

The error message was displayed only on visuals using attributes from the remote dataset.

This message being not self explanatory on Power BI Services, I have tried to download the pbix to check on my local machine: no error. Then, I have tried to use DAX Studio to query the local dataset, calling some attributes from the remote dataset using the XMLA endpoint. I was able to get this error message: 

07/06/2023 15:20:41		
{"error":{"code":"Premium_ASWL_Error","pbi.error":{"code":"Premium_ASWL_Error","parameters":{},"details":[{"code":"Premium_ASWL_Error_Details_Label","detail":{"type":1,"value":"Single sign-on (SSO) credential is not provided which is required to execute the current request. Note that performing a refresh over a dataset with a calculated table or a calculated column which references a Direct Query data source is not supported within the Power BI service."}}]}}}

Détails techniques :
RootActivityId : 0c7dadbb-ecc0-49a1-bdf0-916f743ed4d6
Date (UTC) : 6/7/2023 1:20:40 PM

This error message suggests to get rid of calculated table or calculated columns. Some users in other threads (such as this one) said they were able to fix the issue doing that. However, I think it is not satisfying.

I tried to change the configuration of the dataset on Power BI Services, and found something interesting.

I had someone else from the team taking over the dataset, and configuring the data source credentials (especially for the remote dataset), and it actually worked.

The permissions configured on the remote dataset are the same for the two accounts, so I am stucked in the understanding of this behaviour.

Maybe someone here has an idea?

[ChrisRenlund]

We had the exact same issue - can't say I understand what caused it, but your workaround with taking over both datasets with another user account and reset the credentials for data sources worked for us.
And we could then reclaim both datasets with our service user account again and now it works just fine. 
Thanks!

[gnico]

Note: I tried to create a simple composite model calling the same remote dataset with no calculated table nor calculated columns: I got the same error message. The issue must be account related.

[LucasB]

Did You solve the problem? I have the same issue in my project.

[jaweher899]

Here are my answers to your questions:

1. Incremental Refreshes: For incremental refreshes, you can configure the dataflow to only retrieve the new data since the last refresh, instead of retrieving all the data every time. This can be done by using the "Incremental refresh" feature in Power BI, which allows you to specify the data refresh frequency and the criteria for determining what data is considered new. In your case, you can configure the incremental refresh to refresh daily, and specify the criteria as "date >= yesterday". This way, only the new data added to the data sources since the last refresh will be retrieved.

2. Initial Refresh: To avoid having the dataflow process all the data at once, you can break down the initial refresh into multiple smaller refreshes. For example, you can configure the initial refresh to retrieve only 1 week of data, then run another refresh the next week to retrieve another week of data, and so on, until you have retrieved all the data you need.

3. Staging Dataflows: The staging dataflows should hold only the new data and not the historical data. This is because the staging dataflows are used to prepare and clean the data, and having the historical data in them would increase the refresh time and data storage requirements.

4. Transformation Dataflows: The transformation dataflows should be as granular as necessary to accomplish the desired transformations. In your case, it might be advisable to do the merge and unpivot separately, so you can easily troubleshoot and modify each step of the transformation process if needed.

I hope this helps, and let me know if you have any other questions. Good luck with your dashboard pipeline!

[odumna]

Hi! Team

Any luck? I still has error !!! 

[odumna]

Thank you for your answer.
But I don't get it helpfull.
Because I don't use incremental refreshes and dataflows in my report.
And some time ago it worked fine. Now I get error.
I need to understand the reason of error snd if it can be fixed from power bi service side.

[AMD0791]

That error message really stinks because it's really 2 separate errors to chase!

I've started seeing this error with datasets that are composite models.  I started seeing it a bunch recently when I started taking over datasets for someone who left the company.  I'd take over, set the gateway connections, do the "Edit Credentials" for all the cloud data sources (including the PowerBI datasets), and still get the error.  What seemed to work was to re-publish the same pbix file to a separate workspace temporarily.  As soon as I published the "shadow" version, the original version in the original workspace works fine, and I could then delete the one I had just published.  But it seems like that only works if the temporary publish is done by the owner of the "problem" dataset.

It makes zero sense, but it seems to work.  I've done it with 5-6 reports in the past month.

[wilsonvst]

I had tried to republish PBIX into the main site and my personal workspace, and it works! no longer prompting the error for DirectQuery. Thank you for sharing this great workaround! 

[wojaswojas]

Thank you! It worked for me. Extremely weird, I must admit. But works. Hopefully Microsoft will soon fix this as this is clearly bug