Your question is Reliability in Distributed Failures. 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 ensure service reliability when one component of a distributed system fails?