people · May 18, 2026

ScyllaDB Cut Sprig's Read Latency 4X After Redis and ClickHouse Hit a Wall

Sprig migrated from Redis, ClickHouse, and Postgres to ScyllaDB Cloud on May 5, 2026, achieving 4X lower read latency for attributes. Sprig processed over 1.3 trillion events and 75 billion attributes from 15 billion visitors, outgrowing Postgres due to high write throughput and Redis/ClickHouse due to P99 latencies around 50ms. The migration used a Kafka connector for live writes and a bulk loader for historical backfill with 99.99% data agreement validation. Brendan Cox, Sprig's Senior Staff Engineer, noted ScyllaDB's fast row cache and Materialized Views enabled 500-microsecond mean latencies.