The only potential issue that springs immediately to mind would be vapor escaping through the sleeve in the lid (though, at this point, is anybody even going to notice a little extra evaporation in the face of all the deliberate and ongoing-accidental water releases?), and the possibility of the measuring rod 'binding' if it somehow ends up tilted too far from vertical and placing excessive force on just a couple of contact points (which would quite possibly cause the slip sleeve to bite into the rod and keep it fixed in position even as the water level changes). I'm sure that some clever mechanical engineer has a design for a superior leak-resistant and low-friction slip sleeve; but I don't know the details of such a beast.
Aside from that, though, it was a perfectly serious suggestion. Materials cost, per tank, is peanuts, float-type sensors are fully compatible with electronic instrumentation, if desired; but also work totally passively, and the failure mode still allows you to track state from a safe distance with a clipboard minion and some binoculars (unlike the failure modes of ultrasonic rangefinders, photointerrupters, or similar widgets, which might stop responding or start sending back dodgy numbers, with no ability to verify except by sending somebody into the tank farm to check it out.
I wouldn't want to be the lucky guy who gets to stand on the roof of the shoddily-built radiation-goo tank and retrofit a sensor sleeve; but including it in the design of new tanks wouldn't be difficult.