Your question is Debugging Distributed Production Incidents. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How do you approach debugging a production incident in a distributed system?