red feature flags
Feature flags are such a simple concept, and when done simply (like they should be), they work well. Really well.
Flags create more work for you and your team.
Once you build complexity around them, or fail to manage/kill them, they get messy quick and cause headaches. Teams are notoriously terrible at getting rid of them, too.
Building multiple features behind one flag? Have fun.
Not cleaning up old flags? Delaying rollouts? Enjoy complexity. Enjoy tech-debt.
Not managing the flags in a big org? Often a big bill.
Also, if you're managing feature access long-term through flags: You've now got code that's misleading and permanent; stop doing it. Use something specifically made for it. Yes, it may seem flag-y, but then it's specifically for it.
There are some good solutions out there, if you're happy to manage the flags properly. PostHog comes to mind. They charge per eval, but their free tier is super generous. There are some free tools out there, too. But then you have even more to manage.
Flags should stay out of the way. They should work for you and cost you almost no time. Yet, they're often a lot of hassle.