I personally don't understand this. Say your team owns 2 or 3 interconnected services. A customer reports an issue. How do you go about tracing the route cause?

I personally don't understand this. Say your team owns 2 or 3 interconnected services. A customer reports an issue. How do you go about tracing the route cause? You don't know which server services a request so now you need to fetch logs from all your servers. So now you go and retrieve local logs from all of your servers to your local dev box (which could be considerable size given the size of your infrastructure and request volume) and then try and grep them? What if you need low level wire and request logs to debug the issue? After doing this enough times you are probably going to wish you had centralized and searchable log infrastructure.