System design interview question
Design a Multi-Region Event Streaming Platform with Real-Time Analytics (like Kafka + DataDog Events)
Design a distributed event streaming and analytics platform that ingests events from millions of IoT devices and application servers across 3 geographic regions (US, EU, APAC), processes 500K events/second at peak, and serves real-time dashboards with sub-second latency. The system must guarantee at-least-once delivery semantics, support 30-day event retention with configurable TTL, handle 50TB of new event data daily, and provide low-latency querying for time-series aggregations (percentiles, cardinality estimation) across arbitrary event dimensions. Candidates should design the ingestion layer, storage strategy, query engine, and cross-region replication mechanism while accounting for network partitions, failover scenarios, and cost optimization.
Practice this question step by step with an AI interviewer that reviews your requirements, your architecture and your trade-offs, and scores the design.