Because it's related to your comment, this is a good a place to ask something I've been wondering for a long time:
Is there a word or concept (in English, or in any language) for an ostensibly straightforward and common sense solution to a real problem, whose implementation entails creating a solution which is far larger in scale than the original problem ever was?
I mean something distinct from concepts like perverse incentives (the cobra effect, for example). I mean something like what you're talking about above. But a more illustrative, completely made-up example might be a law that was meant to make people pay their traffic tickets. Imagine that, through a series of steps which each seemed reasonable at the time, implementing this law ended up creating a global, keyhole satellite surveillance network over the entire surface of the country, with massive bunkers of computers monitoring for traffic micro-violations (".02 kmh over the speed limit, that'll be 2 cents") in real time, then notifying people with a tracking bracelet that everyone was legally required to wear, because otherwise it wouldn't be possible to enforce the law we all agreed made sense at the time. And, to ensure the tickets could be delivered in a timely manner, the banks and communications companies had to support a bunch of government technology and oversight, and car companies had to build some tech into their automobiles to shut the off if you didn't pay.
It becomes dystopian really quickly, but I don't think it has to.
This silly example is just meant to illustrate what I mean: is there a name for decisions whose implementation seems straightforward, but which, step by step, create a much bigger apparatus than you originally intended?
I don't think this is entirely theoretical: there are plenty of real world examples, but I won't name them because they would probably distract from the question. Even in software development, you probably run into a situation where some product person says "can't we show this property over here?" and you say "sure, but we'd have to spend a year rewriting our entire architecture first".