Whenever you're ready, walk me through how you'd approach this.
How would you evaluate and monitor a large-scale distributed system for failure modes and trade-offs?
Describe a practical approach covering service-level objectives, telemetry, dependency failures, capacity limits, consistency and availability trade-offs, incident detection, and validation through testing. Explain how you would distinguish infrastructure failures from application defects and use observability data to prioritize improvements.