They have not seen our legacy code base.
The moral of too many of my stories is if something is bothering you, track it first before trying to fix it. Because it's easier to get other people to help police a graph than a series of unix commands. And for everyone else you'll get regressions when you're focused on some other priority. Data points like "sometime in the last month" are pretty hard to nail down on a large codebase, especially if it's not a monolith. Last Tuesday around noon is pretty specific.