Morning baking lists for a bakery chain

Your bakery chain has 40 shops. Every morning, each shop's bakers bake about 60 products from a list. Today each manager guesses the amounts. Unsold bread is thrown away at closing, and empty shelves lose sales.

The data team has trained a model that predicts one shop's sales of one product for the next day. It reads past daily sales, the day of the week, holidays and the weather forecast. Shops close at 20:00, and their tills upload the day's sales by 21:00. Now and then a shop's upload is hours late. A weather service publishes the next day's forecast at 22:00. Bakers start at 04:00 and need their list then.

When does the forecast run, what does it read, and what list does a baker get when something goes wrong?