How to determine compute for fabric capacity?

[AnmolGan81]

Hi, we would like to understand how can we determine compute for a fabric capacity if are migrating from Synapse Analytics where we have a specific DWU, what capcity is equivalent to that, considering we too have Azure Analysis Services that are used to refresh the models, and similarly if we have data on , SQL Server on Azure VM, Azure SQL Database or On-Premise SQL Server, how can we determine which fabric capacity to chosse is there any chart or differences provided by Microsoft around this ask?

[v-kpoloju-msft]

Hi @AnmolGan81,
Thank you for reaching out to the Microsoft fabric community forum.

I understand you are trying to figure out what Fabric capacity would be the right fit when migrating from your current setup (Synapse DWUs, Azure Analysis Services, SQL Server, etc.).

At the moment, there is not a one-to-one mapping between DWUs/DTUs/QPUs and Fabric Capacity Units (CUs). Fabric works a bit differently instead of fixed units like DWUs, you size based on factors such as:

• Dataset size and refresh frequency
• Query concurrency (how many users are active at the same time)
• Memory requirements for your workloads

Please refer these links:
https://learn.microsoft.com/en-us/fabric/enterprise/powerbi/aas-pbi-migration-overview

https://learn.microsoft.com/en-us/fabric/admin/capacity-settings?tabs=power-bi-premium

https://learn.microsoft.com/en-in/fabric/enterprise/licenses#capacity-skus

https://azure.microsoft.com/en-in/pricing/calculator/

Next Steps: Review your current workloads (e.g., Synapse DWU usage, AAS model sizes, refresh times). Try running them in a Fabric trial capacity (F64) to benchmark performance. Use the pricing calculator to estimate cost and capacity needs. Scale up or down depending on your refresh performance and user concurrency requirements.

While there is no official “DWU-to-Fabric” conversion chart, the best way to determine the right Fabric capacity is to test your workloads in Fabric and then choose the closest F SKU that supports your model sizes and concurrency.

Hope this helps clarify things and let me know what you find after giving these steps a try happy to help you investigate this further.

Thank you for using the Microsoft Fabric Community Forum.

[AnmolGan81]

Thanks for the update, but we cannot switch to F64 as it would increase more cost then expected on the purchase capacity, and we are not sure if it will work, if we have some idea on which capacity to be in as compared to synapse then we might be able to provide justification and swithc it.

Will check the links shared.

[v-kpoloju-msft · Accepted solution]

Hi @AnmolGan81,

Thanks for clarifying. I completely understand your concern around the additional cost of moving to an F64 capacity without being sure if it is the right fit. At this point, there is not a one-to-one equivalent between Synapse DWUs and Fabric capacities. The right capacity depends on multiple factors like the size of your models, refresh frequency, number of concurrent users, and workloads (data engineering, real-time analytics, reporting, etc.).

Please go through the below mentioned steps those might help to resolve the issue:

• Start with a lower capacity tier (like F4/F8) and monitor performance using Fabric Capacity Metrics App to see if it meets your needs.
• Use that data to justify scaling up (only if needed), so you are not over-provisioning from the start.
• You can also use the Fabric pricing calculator to estimate costs and compare them against your current Synapse DWU spend.

This way, you will have a clearer view of what capacity level is best aligned with your usage and budget, instead of jumping directly to a higher tier.

Hope this helps.
Thank you for using the Microsoft Fabric Community Forum.

[AnmolGan81]

Thanks for the response, will try this maybe till the time Microsoft releases something as a pre-req guide when migrating from synapse to fabric or any other system to fabric.