Connection Pooling

Category: infrastructure

A cache of database connections maintained so that connections can be reused for future requests.

Opening a new connection to a database is expensive and slow. Connection pooling keeps 50 connections "alive" in a bucket. When a new web request comes in, it grabs an existing connection. When it’s done, it puts it back. It saves immense latency overhead.

Common Examples

• We tuned our connection pooling settings to support high-frequency quoting traffic, which significantly lowered the load on our primary database.
• Properly sizing your connection pooling is a critical performance tuning task, as too few connections stall, and too many crash the database.