Dear Fabric Team,
I am currently using an F16 capacity, costing over $2,500 USD per month, to host a 2.5GB Power BI semantic model. However, I am experiencing a significant issue where the model refresh fails due to resource limitations. The error message is as follows:
"Resource Governing: This operation was canceled because there wasn't enough memory to finish running it. Either reduce the memory footprint of your dataset by doing things such as limiting the amount of imported data, or if using Power BI Premium, increase the memory of the Premium capacity where this dataset is hosted. More details: consumed memory 2865 MB, memory limit 2782 MB, database size before command execution 2337 MB."
Despite the premium cost of this capacity, the model is not functioning as expected, and the memory constraints are causing substantial disruptions to our operations. It is unreasonable that an F16 SKU cannot handle a dataset of this size, especially considering the level of investment in this service.
Additionally, with the same level of investment, I can get a 50GB capacity from Azure Analysis Services. This raises concerns about the value proposition and efficiency of the current setup with Fabric, as it does not seem to provide comparable scalability and performance for a significantly higher cost.
I would appreciate an immediate investigation into this issue, along with a solution that will allow us to effectively utilize the capacity we have paid for. If this cannot be resolved with the current capacity, I would like to understand what alternatives or adjustments are available to ensure smooth performance without compromising our datasets.
Thank you for your prompt attention to this matter.
Best Regards
TK
