Predicting phone sales with Python and SARIMAX or other

I am trying to predict sales for phones and there are several factors.

1. launch day can amount to sales for several months up to 6 months

2. seasonality plays a huge role, I have about 5 years of data

3. sometimes sales stop because the phone is out of stock (but otherwise there is demand)

4. specific models have usually short life span so there is not much data, working mostly on 2-6 months data

I have been using SARIMAX to incorporate 2 and 4 together quite nicely. But I'm unable to resolve 1 and 3 for my data.

What would you suggest?

I tried reading SARIMAX documentation and annoying ChatGPT for solutions.