And if people can pick their own vegetables, or pick them and sell/use them the same day, extremely locally, then you can select for taste instead of for transportation resilience.
It's well beyond feel-good.
They. Are. Everywhere.
They routine infest indoor houseplants, even in places you wouldn't expect. They'd find their way into a roof garden. And because most predators require a wider range than most prey, there may not be as many natural predators available.
Commercial greenhouse growers definitely try to isolate their plants from invasion, but still rely on pest control when the inevitable happens.
Predatory mites work fine if other mites are the only problem. They aren't particularly expensive either.
But yes, you can't use natural predators. You basically need to introduce them, let them do their jobs, and then most of the predators will die once the problem is controlled for.
But give its a $50 problem, its not a very expensive problem and on par with pesticide costs.
So, quite frankly, your statement that the efficient distribution of water is fertilizer is more is just wrong.
I'm not saying the tech is perfect or that it can't improve, but if even a fraction of the research dollars were put into this as is put into traditional/industrial research and development, I think aquaponics (especially when designed vertically) will outpace it easily.
I will note that this isn't true for many staple crops like corn, soy, or wheat. But that's not the point. You would need to compare specialty crop production, not commodity crop production.