Dynamically estimating and scaling Postgres’ working set size

Neon has introduced an automatic scaling feature for its Postgres instances to keep the working set in memory. The working set size is estimated using HyperLogLog (HLL), a probabilistic algorithm that estimates the cardinality of a set with high accuracy and minimal memory usage. By dynamically estimating and adjusting the working set size, Neon can provide significant speed-ups for real-world workloads while ensuring predictable performance. The new feature is expected to be more cost-efficient than fixed-size nodes with the same performance.