The obvious solution is not always the right one.
When we hit a problem in OCI where KMS keys were refusing to delete; the obvious move was to force delete them. Try harder. Debug the script. Fix the error.
We tried that. It kept failing.
So we stopped and asked a different question.
What if we stop trying to delete them where they are; and move them somewhere else first?
We wrote a script that moves the keys to a separate isolated compartment. No active resources. No dependencies. Clean environment. Then we delete them there.
It worked immediately.
Same problem. Different angle. Completely different result.
This is something I keep learning over and over in cloud engineering. When something is not working; the answer is rarely to push harder in the same direction. Sometimes you need to zoom out; question your assumptions and ask what else is possible.
The cloud does not always behave the way you expect. Neither do real problems.
The engineers who solve hard problems fastest are not the ones who know the most commands. They are the ones who think differently when the obvious path is blocked.
When the door is locked; look for the window.