I would be even more cautious. An in-place upgrade may be an interesting training exercise, but has considerable risks for anything else.

If you are doing an in-place upgrade your planning should include the scenario of the upgrade failing part-way through - these things can happen. For the OS upgrade this could mean your server is unusable and needs to be restored to state prior to the OS upgrade starting. For the SQL upgrade this could mean SQL is unusable and your server needs to be restored to a state prior to the SQL upgrade starting. If you are updating a production machine you should test these scenarios to make sure the restore process will work.

IMHO the risks and work needed to do an in-place upgrade are not worth it for a production machine. You should always aim to do a side-by-side upgrade for production, and anything else that your organisation relies on.