Quick answer: Weigh the bug's impact against the cost and risk of fixing it, and consciously leave it alone when the fix is not worth the return. Deciding not to fix a bug is a legitimate, necessary engineering decision, your time is finite and some bugs genuinely are not worth it.
It feels wrong to leave a known bug unfixed, but trying to fix every bug is neither possible nor wise. Your time is finite, and some fixes cost more, in effort, in risk of introducing new bugs, than the bug itself costs your players. Deciding deliberately that a bug is not worth fixing is a legitimate engineering decision, not a failure. The skill is making that call honestly, based on impact versus cost, rather than either fixing everything or ignoring things by accident.
Some Bugs Genuinely Are Not Worth It
Accept the premise first: it is correct, not lazy, to leave certain bugs unfixed. A cosmetic glitch that two players will ever notice, in a rarely-visited corner of the game, fixed at the cost of hours you could spend on a bug hitting thousands, is a bad trade. Engineering is allocation under constraints, and choosing not to fix low-value bugs is how you free your finite time for high-value ones. A backlog with consciously-deferred bugs in it is healthy, not negligent.
The alternative, treating every bug as mandatory, leads to either burnout or to high-impact bugs waiting while you polish trivial ones. Giving yourself explicit permission to not fix some bugs is what makes good prioritization possible at all.
Weigh Impact Against Cost and Risk
The decision is a cost-benefit calculation. On the benefit side: how many players does the bug affect and how badly, occurrence counts and severity give you this. On the cost side: how much effort the fix takes, and crucially, how much risk it carries of introducing a new bug. A low-impact bug that needs a risky change to a core system is a clear not-worth-it, the potential downside dwarfs the upside. A low-impact bug that is also a trivial safe fix might be worth doing anyway.
Bugnet's occurrence data grounds the benefit side of this in reality, you can see exactly how few players a bug actually affects, which often reveals that an annoying-seeming bug is genuinely low-impact and not worth a risky fix. Pairing real reach data with an honest assessment of fix effort and risk turns 'should I fix this?' into a defensible decision rather than a guilty guess.
Decide Consciously and Document It
The important thing is that not-fixing is a decision, not an accident. When you decide a bug is not worth fixing, mark it as such in your tracker, deferred, or wont-fix with a reason, so the decision is recorded and the bug is not simply forgotten. This keeps the call honest (you can revisit it if the bug's impact grows) and keeps your backlog meaningful (every bug is either actionable or consciously set aside, never just lost).
Documenting the reason also helps with players. If someone reports a bug you have decided not to fix, you can respond honestly: it is real, but it affects few players and the fix carries risk, so it is logged in case that changes. That respects the reporter while standing by a reasonable decision. Deciding when a bug is not worth fixing, weighing impact against cost and risk, recording the call, and being honest about it, is a core part of running a game well, not a corner you are cutting.
Not every bug deserves a fix. Weigh impact against cost and risk, decide consciously, and record the call, don't fix by reflex or ignore by accident.