Azure Bill on a Fabric Capacity · Learn Microsoft Fabric

Azure Bill on a Fabric Capacity

Hi everyone!

I’m starting to delve into this new world of Fabric, understanding the architecture, pricing model, etc. I have an initial question about Azure pricing in Fabric that I don’t fully understand. If I have my capacity deployed as "Pay-as-you-go," will Azure charge me hourly for all the "CUs" (capacity units) in my SKU as long as my capacity is "on," regardless of whether I'm using the CUs or not?

Or does Azure only charge me based on the actual consumption and utilization of the available "CUs" in my SKU?

Another question I have about Azure's pricing model in Fabric concerns the use of Spark. Reading through the documentation, it’s not entirely clear to me whether the use of Spark incurs an additional cost beyond the reserved CU consumption for my capacity. If Spark consumes and utilizes the CUs, which have vCores and memory allocated for their use, why would it generate an additional charge when CUs in the capacity are already being charged?

Yesterday, I ran a test by creating an F2 to start interacting and see Azure bill when using Spark and interacting with OneLake and a Lakehouse. Looking at the costs, besides the CU compute consumption, I also see a specific consumption item called "Spark Memory Optimized Capacity Usage CU."

So now I’m wondering, in the first two lines, what does "Compute Pool Capacity Usage CU" refer to? Is it a fixed cost for having the capacity turned on?

Thanks for reading.

1 comment