Software takes on entirely new levels of complexity when it’s being deployed onto custom hardware environment that hasn’t been used hundreds of thousands of times before.
Frankly, every time I have a bit of hardware that doesn’t interact correctly with the world, I have a minor fit of panic: “How will I be able to tell what’s broken? How will I know when it’s fixed? How can I hope to keep this thing working in the future? How can I have any certainty at all that my fix for one thing didn’t break all the timing requirements for a different feature? This box is a claptrap of side effects and there’s no way I can wrap my head around that much complexity in any reasonable amount of time! Aagh!“
The answer is to address sustainability issues early on when developing a software-in-custom-hardware product. For you own peace of mind, put these things in place before too much is implemented.