Your question is Designing for Traffic Spikes. 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 would you design a distributed system to handle a sudden spike in traffic?
Explain how you would detect the spike, preserve availability and latency, scale compute and data layers, and degrade gracefully when capacity is exhausted. Discuss where machine learning could improve forecasting or routing, how you would evaluate it, and how you would monitor failures in production.