Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 

Earn the coveted Fabric Analytics Engineer certification. 100% off your exam for a limited time only!

Reply
dataX
Frequent Visitor

Power BI architecture in the context of Azure

Dear Power BI community,

                                              

I have a question with you experts about how to implement a Power BI architecture in the context of Azure Cloud and POwer BI service.

 

Couple months ago, we started a POC with Power BI, Power BI service connecting to Azure database and data being loaded in Power BI itself (input mode not directquery) and the Power Bi service .All worked fine and dandy as we had little data to start with. Now we loaded our full data into the file, the pbix is 180MB. As we did  that we are also starting to see some circles(latency) when a filter is selected between the visualizations and we are starting to doubt this approach. Additionally we got a requirement we have to share this with external users who will be editing the file so started a Group workspace as well with with Pro License.

 

With the data limilations on the Pro account(10 GB) and overloading the data in PBIX is causing a performance hit, we are starting to look into Direct Query. With Direct query data resides on the database, is SQL Azure or SQL data warehouse is capable of getting us rid of the perfomance or we need a SSAS tabular engine? Can you please share your ideas in the Context of Azure and Power BI service if you already have implemented something similar. Thank you in Advance!

 

 

2 REPLIES 2
Eric_Zhang
Employee
Employee

@dataX

 

As per my experience, the import mode usually has better performance than DirectQuery, so I don't think you would get better performance via using DQ mode.

 

180MB seems not quite huge data, how is the lantency like, 1 second, 2, 3 or even more? In your case, the lantecy is probablly caused by network due to Geography factor rather than import mode.

 

AFAIK, there's very limited things we can do to improve the Power BI service performance, I'd suggest you sumbit a support ticketPro users get free support.

 

By the way, if you do think the data size may affect the performance, maybe you can try to split the report into reports or import aggregated data.

I have been trying to understand how large scale datasets are handled.

 

Here is my scenario. I have a db per tenant (customer), and plan to create reports per customer which internally point to the corresponding customer db.

 

Im unclear about how the datasets are handled internally by powerbi service, and what levers are available to me to manage concurent data sets for 10s of customers, as well ensure required latency, e2e from db to powerbi service to powerbi embedded vnode to the web app dashboard.

 

Where can I find this information?

Helpful resources

Announcements
April AMA free

Microsoft Fabric AMA Livestream

Join us Tuesday, April 09, 9:00 – 10:00 AM PST for a live, expert-led Q&A session on all things Microsoft Fabric!

March Fabric Community Update

Fabric Community Update - March 2024

Find out what's new and trending in the Fabric Community.

Top Solution Authors