Learn how to build advanced real-time data analytics systems by exploring a 39-minute technical talk that demonstrates the evolution from traditional Kafka-Flink architectures to a more sophisticated solution using Apache Pulsar, ClickHouse, and StarRocks. Discover how to leverage Pulsar's FunctionMesh and geo-replication capabilities to overcome traditional system limitations, implement efficient data segmentation, and enable flexible data reuse. Explore the integration of ClickHouse for handling large-scale real-time analysis and understand how StarRocks enhances the system by simplifying complex mesh aggregation tasks. Gain valuable insights from five years of practical experience in optimizing data analytics systems, including strategies for scaling, managing data surges, and implementing recovery scenarios.