I think you have the null hypothesis backwards here. Do you have any statistics to show that issues _are not_ dropping through the cracks? Sending a “ping?” message on a ML is generally considered poor netiquette, and even if it were expected here, would make many requesters uncomfortable. That’s one of the fundamental things any tracker does—keeps statistics on and forces every issue to _some_ resolution, even if it’s “will not fix” or “on hold”. Things don’t just peter out and get forgotten like on email threads.

(I have not done the exercise of perusing old email threads to see if I can find issues being dropped on the floor. But I’ve already found several apparent existence proofs. Whether they’re common or rare is a question I can’t answer without doing tedious manual work that is the entire raison d’être of a tracker.)

I wouldn’t dispute that the Linux kernel ML, for the most part, works. But the Linux kernel mailing list is a rather different beast than the potential users of an issue tracker for any other software project I can imagine—the technical acumen expected of contributors is high, quotidian back-and-forth user-assistance exchanges with non-contributors are not tolerated, people are usually expected to offer fixes as working code and not simply prose bug reports or feature requests (except for critical or security issues), and patches and pull requests on the mailing list are dealt with using a distributed version-control system that was purpose-built for the task (though happens to have worked well enough to because the most widely-used DVCS period).