How do you handle high-throughput ingestion while maintaining low-latency search capabilities for Bloomberg stories?
