Get certified for free when you join Fabric Data Days 2026 and dive into Fabric, Power BI, SQL, AI, and other essential data skills.
Join nowData Days is here! Join us now for 60+ days of learning, challenges, and connection. Learn more
Dear Community,
I have a question :
In a Big Data Architecture (Data Lake Store), is connecting Power BI to Azure SQL Datawarehouse through DirectQuery not to heavy to support for it ?
It's not going to saturate the dababase ?
Thank you for your answers.
Solved! Go to Solution.
@Anonymous In my opinion DQ is valueable for extremely large datasets that you can't stick in Azure Analysis Services, if you can - I'd recommend adding the AS layer and having Power BI live connect to that. You'll get much better performance and you won't need to model in the PBI reports. If you are "at scale" where that isn't feasible, you just need to ensure you are tuning your queries in the WH to be as performant as possible because every interaction by a report user will execute a query. So you will need to do some anlysis on expected workload and user base to understand the type of performance/sizing you require for the WH to support internal and external workloads.
@Anonymous In my opinion DQ is valueable for extremely large datasets that you can't stick in Azure Analysis Services, if you can - I'd recommend adding the AS layer and having Power BI live connect to that. You'll get much better performance and you won't need to model in the PBI reports. If you are "at scale" where that isn't feasible, you just need to ensure you are tuning your queries in the WH to be as performant as possible because every interaction by a report user will execute a query. So you will need to do some anlysis on expected workload and user base to understand the type of performance/sizing you require for the WH to support internal and external workloads.
Thank you for your answer !
So, is AS + LiveConnect a suitable solution for a large company with a lot of users ?
@Anonymous Yes, it is from an enterprise reporting standpoint. There are a whole slew of ways in which you can implement solutions, but from an enterprise level AS should be part of your solution to support larger models and needs for enterprise reporting.
Don't miss out on Data Days, June 15 through August 7. Learn Fabric, Power BI, SQL, AI and more.
Check out the May 2026 Power BI update to learn about new features.
| User | Count |
|---|---|
| 16 | |
| 10 | |
| 10 | |
| 8 | |
| 7 |
| User | Count |
|---|---|
| 41 | |
| 36 | |
| 34 | |
| 28 | |
| 20 |