How would you design a real-time data ingestion pipeline for high-volume transactional data for Pismo?
