Tuning the HikariCP Connection Pool

Database connections are expensive to open, so Spring Boot pools them with HikariCP by default. When throughput stalls under load, an undersized or misconfigured pool is a frequent — and invisible — culprit.

The key settings

Most tuning happens through a handful of properties. Boot maps them straight from application.properties.

spring.datasource.hikari.maximum-pool-size=20
spring.datasource.hikari.minimum-idle=5
spring.datasource.hikari.connection-timeout=3000
spring.datasource.hikari.max-lifetime=1800000
spring.datasource.hikari.idle-timeout=600000

Bigger is not better

More connections than the database can serve just moves contention from your app to the database. A common starting formula is connections = (cores * 2) + disks, then measure.

@Bean
public MeterBinder hikariMetrics(HikariDataSource ds) {
    return registry -> ds.setMetricRegistry(registry);
}

Exposing pool metrics lets you watch active vs. idle connections and pending threads waiting for a connection.

Fail fast, don't hang

A short connection-timeout means a request fails quickly when the pool is exhausted, instead of piling up threads that block indefinitely.

Takeaways

• The pool, not the CPU, is often the limit under database-heavy load.
• Oversizing the pool pushes contention onto the database — size deliberately.
• Keep connection-timeout short so exhaustion fails fast and visibly.

Set max-lifetime a little below your database's own connection timeout so the pool retires connections before the server drops them.