Whenever you're ready, walk me through how you'd approach this.
How would you design a highly available service to handle a massive spike in traffic at a company?
Explain how you would clarify the workload, protect dependencies, scale stateless and ML-serving components, and preserve availability when capacity or downstream systems are constrained. Cover traffic admission, graceful degradation, observability, and recovery. State assumptions explicitly rather than relying on unstated product or traffic details.