Estimating the Best SKU for PowerBI in Fabric via Azure

[YuyaoWang]

I have been assigned the task of estimating the cost for using PowerBI in Fabric, specifically for retrieving data via prebuilt connectors from the Celonis platform. However, I’m currently a bit confused about how to proceed.

It seems that the purchase of Fabric is managed through the Azure portal, and I now need to choose the appropriate SKU. Could you please help guide me on how to estimate the best SKU for our use case? Any suggestions or criteria would be highly appreciated.

[AMeyersen · Accepted solution]

There are several factors to consider when calculating total cost.

• number of users and licensing model. Report viewers can use a free Power BI Lincense if you have a F64 capacity or above. Otherwise you need at least a Power BI Pro license for each user. 
• data volume and complexity of transformations. More data and more complex data operations require more computing power -> higher SKU
• frequency of data refresh. If you refresh data infrequently (twice per day or less) you might want to dynamically scale your SKU up for the refresh operation and the scale it down afterwards.

As a rule of thumb: if you have less than ~200 report consumers, you might want to consider buying a Pro license for each user and automatically switch your Fabric capacity off or scale it down whenever the data refresh is finished.

Otherwise start with a F64 capacity (because of free viewer licenses) and monitor the resource consumption closely. Once you have stable usage data you can calculate the capacity you need and use the long term reservation discount if you buy it for one or three years in advance.

[Poojara_D12 · Accepted solution]

Hi @YuyaoWang 

• Start with Fabric Trial: Test workloads with the free trial (F64 SKU) to understand resource needs.

• Key Factors: Consider data volume, refresh frequency, concurrency, and required features.

• Estimate SKUs:

  • F2/F4 for small datasets (<10 GB).
  • F8 for medium datasets (10–100 GB).
  • F16+ for large datasets (>100 GB) or high usage.

• Monitor & Scale: Use Azure Capacity Metrics to monitor usage and adjust the SKU as needed.

• Additional: Ensure Power BI Pro/Premium licenses and verify Celonis connector requirements.

Did I answer your question? Mark my post as a solution, this will help others!

If my response(s) assisted you in any way, don't forget to drop me a "Kudos" 🙂

Kind Regards,
Poojara
Data Analyst | MSBI Developer | Power BI Consultant
Consider Subscribing my YouTube for Beginners/Advance Concepts: https://youtube.com/@biconcepts?si=04iw9SYI2HN80HKS 

[AMeyersen · Accepted solution]

There are several factors to consider when calculating total cost.

• number of users and licensing model. Report viewers can use a free Power BI Lincense if you have a F64 capacity or above. Otherwise you need at least a Power BI Pro license for each user. 
• data volume and complexity of transformations. More data and more complex data operations require more computing power -> higher SKU
• frequency of data refresh. If you refresh data infrequently (twice per day or less) you might want to dynamically scale your SKU up for the refresh operation and the scale it down afterwards.

As a rule of thumb: if you have less than ~200 report consumers, you might want to consider buying a Pro license for each user and automatically switch your Fabric capacity off or scale it down whenever the data refresh is finished.

Otherwise start with a F64 capacity (because of free viewer licenses) and monitor the resource consumption closely. Once you have stable usage data you can calculate the capacity you need and use the long term reservation discount if you buy it for one or three years in advance.