Reliability, Observability and Safety Questions

Encompasses building reliable and safe systems through observability instrumentation and operational practices. Key areas include telemetry design with metrics logs and traces, alerting and escalation policies, service level objectives and service level agreements and how to use error budgets, runbooks and incident response processes, postmortem culture and continuous improvement, graceful degradation and fallback strategies, retry and idempotency patterns, capacity planning and autoscaling, canary deployments and progressive rollouts, and domain specific considerations such as monitoring model performance or output quality for large language model systems. Candidates should reason about trade offs between cost and reliability, instrumentation coverage, detection latency, and how to measure and improve operational readiness.