Can you explain more about this? Most of the runbooks at my current job exist because it was too painful to leave them unwritten, and get used frequently.
Runbooks are, in my experience at least, generally sticking plasters over systems that are convoluted, unreliable, or otherwise broken. They're a symptom of teams that are running so hot that they can only put out fires, not prevent them. Again, just my experience, but whenever I see an extensive list of runbooks I run a mile.
No, many older systems work perfectly well and are sensible and intuitive. Vertical scaling a single box or HA pair, for instance, is generally easier than working with distributed systems, for instance, and is perfect for many (not all) use cases. The idea that recent stuff isn't terrible is.. well, your anecdote against mine, I guess.
