Your question is Reliability and Fault Tolerance Patterns. 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).
What architectural patterns do you use to ensure reliability and fault tolerance in multi-tiered cloud services?
Explain how you would structure service boundaries, dependencies, data stores, and deployment units to tolerate component, zone, and region failures. Address graceful degradation, observability, recovery, and how reliability decisions affect ML inference or model-serving components where applicable.