Kubernetes Operators: Are they really a boilerplate killer, or do they just add *more* complexity?

ive been seeing a lot more talk about Kubernetes Operators lately, especially for managing stateful applications like databases or message queues. the idea of automating complex operational tasks like upgrades, backups, and scaling sounds fantastic, especially for reducing boilerplate. but im also wondering if they just shift the complexity into the operator itself. for those using them, are they truly simplifying things in the long run, or are they introducing a new layer of abstraction thats hard to debug? whats your experience?