I get it. It does have a learning curve. This being said, I would argue that without selinux Linux can’t really be meaningfully secure. It’s worth learning. Seljnux exits elsewhere too. I deploy Debian with selinux and it works well there as well.
The problem with SELinux is that everyone rushed to push it out, alongside packages affected by it without support for it. So it was a crapshoot whether or not you'd have something working each time. That is better now, but was initially a colossal pain in the ass for about five years or so.
What put me off selinux is that the officially documented way of generating a new policy is to run a service unconfined, and then generating the policy from its behaviour. This is backwards on so many levels... In contrast policy-based admission control in kubernetes is a delight to use, and creating new policies is actually doable outside of a lab.
You could preemptively write the policy if you know the context and policies you want to apply. I just don’t think it’s worth the time when you can generate a policy with two commands.
Fair. But audit2allow makes it really easy to add support for apps without policies. For custom in-house apps I use this to spit out some nice policies that can be rolled out.