ah and that’s why 2x10kW instead of 1x20kW. ;-)
A 10kW inverter is clearly too small.
A 15kW inverter might be a bit too small, but could fit.
A 20kW inverter costs maybe 200€ more than a 15kW.
But well, that could/should be simulated again.
I always find 2 inverters disadvantageous in terms of space consumption, costs, efficiency, failure risk, management, ...
so again. Half faces south. That’s 10kWp. When there is full sun at noon, a maximum of 10kW with 70% power limitation comes from the roof.
If the north roof generates yield, whereas the north roof will never deliver 10kW, the south accordingly delivers significantly less.
I am still inclined to agree that the 10kW inverter could (!) be too small for the system overall. But more than 15 is definitely wrongly designed. You might get problems with too low string voltage.
Two inverters have the advantage of failure safety (why do you claim the opposite - strange) and you can register two separate systems, which can have advantages in feed-in tariffs and/or taxes. Also, one of the systems could be designed as a full feed-in system, etc. - you have to see if such a constellation could be worthwhile.
I would present the roof in a well-known forum.