System design interview question
Design a Distributed Event Streaming and Real-Time Stream Processing Platform for Enterprise Data Pipelines (Like Kafka + Flink at Scale)
Design a distributed event streaming platform that ingests, stores, and processes real-time data streams for enterprise customers at petabyte scale. The system must handle 10M events/second across 500+ enterprise tenants with strict ordering guarantees within partitions, exactly-once semantics for critical applications, and sub-500ms end-to-end latency for stream processing jobs. Support multi-region deployment with failover, consumer lag tracking, schema evolution, and dynamic scaling of processing topologies. The platform must maintain data durability across multiple availability zones, support both pub-sub and stream processing workloads simultaneously, and provide isolation between tenants while optimizing for infrastructure cost.
Practice this question step by step with an AI interviewer that reviews your requirements, your architecture and your trade-offs, and scores the design.