I feel you on the pain of coding. It's a big reason I shifted to management. And it turns out, after a few years of getting my feet under me, I quite enjoy it.
The reality is that in many businesses, the job of software engineering past the entry level is a lot more than just coding.
This isn't a dynamic that is limited to software engineering. I'm renovating my house, and it has been extremely frustrating at times to work with my architect. It occurs to me that those frustrations are typically around the aspects of her job that aren't design oriented, and I can see how those other things probably dominate her work.
Coding is very creative, and I suspect that's what drives many of us to get into the career. But the deeper you get, the more you find that the highest value things you can do from a business are bridging your creative work to all of the systems and processes. This is how you can create a business that harnesses the work of hundreds of producers to meet coherent goals. And unfortunately, this can't be completely externalized to all of the administrative people, like PMs and managers.
What you wrote also speaks to another dynamic, which is that over time, there is a divergence between the platform and the real business needs. More and more time gets spent working around that mismatch. There is no silver bullet solution for this. You can shrink the platform (e.g. microservices) but then you will feel the pain in increasingly complex integration and operations.
But the good news is that platforms are constantly improving, and I think we're getting closer to the point where we stop solving all the wrong problems. For instance, server-side rendering is having a renaissance, getting us away from all of the emergent complexity of client-side frameworks. Tools like Temporal, Airplane.dev, Trigger.dev, Patterns.app, etc. are creating platforms for writing operations and automation tooling with far less operational overhead.