Fabric Capacity Metrics

Hi everyone.

I have been going back and forth with my Dev team on working out some issues with our capacity for the PBI Service. I have a very large report that runs every morning. This report references a large number of dataflows and other queries, but also is connected to other PBI reports as well. I am using the Fabric Capacity Metrics app to watch the performance, but I’m looking for some better insights as to what matters most with these metrics.

For example, this report shows as 28GB for “Item Size”. Can you confirm if this metric really matters all that much with a PBI report? The whole intent of PBI is big reporting/big data, correct? This report is used as a primary database for other reports across the org as well. I’m trying to get a better understanding for when the Item Size really matters. The actual file size is 3GB on desktop - so I assume the “28GB” is include all of the data that is brought in PLUS the data in the other referenced reports? Can someone confirm?

Should I be looking at CU(s), Duration(s), Item Size, etc.? What matters most?

Thank you in advance for your help!