Whenever you're ready, walk me through how you'd approach this.
How would you architect a system for high availability?
Describe a practical architecture for a production data pipeline, including ingestion, processing, storage, orchestration, and serving. Explain how you would handle failures across nodes, availability zones, and regions, while preserving data consistency and supporting recovery, monitoring, and backfills. Include technology choices, deployment topology, and the tradeoffs between availability, consistency, cost, and operational complexity.