Recently, we encountered an issue where as soon as the free MAR period ended, the Usage page showed a massive spike that was unrelated to any new tables being added. It appears that the Usage tab doesn't actually show the daily rows loaded effectively. As you can see in my screenshots, the real data loaded for the days shown during the free MAR period were in the hundreds of millions, but the Usage page doesn't show that. This made for a great surprise the day we started paying for MAR. The free MAR period is meant to show what you would be seeing during the paid period, and not showing certain tables that are loading is both ambiguous and obfuscating with respect to MAR.

Ambiguous - It is can be very difficult to know which tables to turn off if you don't see that they are loading daily via the Usage tab. The Loaded Rows column doesn't show tables, and going through a massive connector e.g. Netsuite to see which tables are loading the most rows is time intensive. A clearly defined and complete list of what is loading daily is the best solution to understand MAR usage, free period or not.

Obfuscating - to not show all the tables loaded by a connector during the free period (and they are definitely loading daily) is to essentially hide a future cost. This feels like a bait and switch - "yea your MAR usage is soooooooo low....oh wait, there's these other ones we weren't showing you, it's actually very high. That'll be $10,000 please". It should be shown as part of the free MAR period to provide clarity and truth that translates directly to the paid MAR period.

I hope this changes. This is coming very much from a place of truth where we had our netsuite connector jump from a few hundred thousand rows during the free MAR period (after we turned off tables that were showing as loading millions of rows daily post-initial sync) to 383MM rows on the first day of paid MAR. That is 2 months of our annual MAR budget in one night, all because the free period was not accurately showing the amount of MAR loaded daily. It should have been.