Increasing Capacity

[Hussein_charif]

Hello all,

i wanted to get professional opinion on upgrading our Capacity. so, here is what we currently have :

we are currently on an F16 capacity with 10 users, we have around 30 dataflows with small amount of data extracted from Business central that run 4 times a day(takes 15 minutes in total), we have 2 dataflows with medium load, that run 6 times a day(each dataflow takes 40minutes approximately), we have 3 pipelines that run every 3 minutes, those pipelines do an incremental refresh getting data from dataverse, each run takes 30seconds-1min, we have 6 reports, all direct query, and multiple paginated reports created and extracted daily by multiple users, we currently have in total in our fact table around 70 columns, nearing 30 million records increasing daily by ~ 1.8m to 2m. the average CU usage is 50-70%, and sometimes when multiple paginated reports are being created and extracted, the throttling % goes over the 100% limit.

we have over 30 new users pending to have access on the reports and paginated reports.

i was aiming to increase the capacity to F64, for multiple reasons being reducing capacity usage; having some breathing space for the capacity with the CPU/memory increase, and the new and existing users wont need a license since on F64 users with viewer access can view the reports and paginated reports without a license.

is that a good option to go with? would love to hear any professional's perspective on this!

thank you in advance!

[cengizhanarslan]

Benefits of moving to F64:

1. Massive headroom increase

Compared to F16, F64 gives you:

• 4× CPU

• 4× memory

• Much higher concurrency tolerance

This directly addresses:

• Throttling during concurrent paginated exports

• Overlaps between refresh + pipelines + user activity

2. License simplification

On F64:

• Viewers do not need Pro or PPU licenses

• This is ideal with 30+ new users pending

• Especially valuable if users consume:

  • Reports

  • Paginated reports

  • Subscriptions

This alone can justify the upgrade financially.

3. Growth protection

Given your data growth rate:

• 2M rows/day ≈ 60M rows/month

• You’ll likely double your fact size in months, not years

F64 gives you runway, not just relief.

[djurecicK2]

F64 is a great option- do you have the budget for it?

[Hussein_charif]

Hi @djurecicK2 , we do yes

[djurecicK2]

Go for it! With that level of increase in record volume, you also might look at using a Direct Lake model.

https://learn.microsoft.com/en-us/fabric/fundamentals/direct-lake-overview

[Hussein_charif]

actually about that, i tried using direct lake model, but i felt it had the same performance as a directquery, any idea why?

[djurecicK2]

Take a look at this video- it is from last year so there have been some changes, but overall is still very good: https://community.fabric.microsoft.com/t5/Power-BI-Community-Blog/Optimizing-Direct-Lake-Mode-in-Mic...

[Hussein_charif]

thanks!