Your question is Root Cause in AWS Incident. 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 are given access to an AWS environment with an ongoing incident: find what is going wrong, identify the root cause, which signals/logs/monitoring you would check, the immediate mitigation, and long-term fixes.
Asked in the incident reporting stage. Treat this as a hands-on production investigation rather than a purely conceptual design exercise. Explain how you would distinguish an application, infrastructure, data, model-serving, or dependency failure, and how you would communicate confidence as evidence accumulates.