Your question is Isolate Variables in Production Outage. 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).
You’re responding to a broken production system and need to determine what changed, what is actually failing, and how to narrow the problem space quickly without making the incident worse.
What is your methodology for isolating variables when troubleshooting a broken production system?