This is my setup

I'd expect A and B production to be pretty similar. And C times 5/9 should be about the same as D. That's not really what I'm getting.

Snow, of course, makes this more complicated.

These are the numbers I'm being told from my installer/Tesla.

The installer's theory is that:
(MPPT B) is quite a bit lower than (MPPT A) for the same array size on the same south facing side of the roof but could be related snow melt on the peak prior to snow melt on the lower roof. This theory seems to be the case on the north facing roof, with the lower array being lower than would be expected.

Seems kind of odd in November though. I can't say the exact date that tree had no leaves, but it's not exactly hanging over the panels. When the sun gets to a low angle, production dies regardless of the tree, so I don't think it's making a huge difference, particularly in the winter. Plus periods of snow are marked by 0 or near 0 production. They're not that hard to spot.

I want to say that a 20-30% difference in A's production vs B's production probably means something's wrong. Any thoughts would be welcome.

I also don't know why their numbers are so far off of what Home Assistant is showing (other than watts vs kilowatts).