Integrating Autoscaling into F SKUs for Enhanced Flexibility and Efficiency

[joshua_wallace]

Currently, Power BI's P SKUs offer autoscaling capabilities, a feature not extended to F SKUs. This proposal suggests the introduction of autoscaling functionality for F SKUs within the Azure Portal. The current model permits manual scaling, allowing for upgrades or downgrades at any given time. However, this approach lacks the finesse needed for handling transient spikes in demand efficiently. Rather than necessitating a full doubling of capacity, which can be both excessive and cost-inefficient for minor demand increases, the implementation of an autoscaling feature would allow for a more nuanced scaling approach. Specifically, it would be beneficial for users to have the option to automatically scale their capacity by adding a precise number of Compute Units (CUs) to manage daily fluctuations in demand. This feature would not only optimize resource utilization but also enhance cost-effectiveness for users requiring temporary increases in capacity without committing to a permanent scale-up.

[msmith6]

It would be good if we could add additional CU to SKUs, similar to how we can add vcores in SQL Server. For example if I have a workload that is just tipping over F64, my options are either somehow split it into a F2 and F64 (or similar), which means the items in the F2 capacity won't have the same rich feature set as the F64 hosted as some things require a min of F64. The other option is to double spend and jump to a F128 which might not make sense if capacity is just tipping over.

[v-beeras]

Customers want an autoscaling option for the Fabric capacity in Power BI, similar to the premium capacity, rather than using the scale-out option in Azure.

[oaladekomo]

Thank you for your suggestion. We are currently evaluating options to bring autoscale to Fabric capacities and will share details with the community once we have them.