socketTimeout uses double the value?

[kncesarini]

Not sure if i should file this as an issue or i'm just misunderstanding something, so figured I'd start here. The problem I'm having is that whatever value I specify for socketTimeout (below 15000), the timeout will instead be after twice that time, in this example 30 seconds.

If i set socketTimeout to 60 seconds, it will time out after 120 seconds, and so on. The query i'm performing so test is just a SELECT pg_sleep(:seconds), but i see the same behavior with regular queries in our real environments.

This profile is used when running the application in AWS/EKS

spring:
  datasource:
    url: ${AWS_WRAPPED_DB_URL}
    username: ${DB_USER}
    password: ${DB_PASSWORD}
    driver-class-name: "software.amazon.jdbc.Driver"
    hikari:
      exception-override-class-name: "software.amazon.jdbc.util.HikariCPSQLException"
      data-source-properties:
        wrapperPlugins: "driverMetaData,auroraConnectionTracker,failover2,efm2"
        wrapperDialect: "aurora-pg"
        connectTimeout: 10000
        socketTimeout: 15000
        monitoring-connectTimeout: 5000
        monitoring-socketTimeout: 5000
        wrapperDriverName: "PostgreSQL JDBC Driver"
        max-lifetime: ${DB_CONNECTION_MAX_LIFETIME:1800000}
        idle-timeout: ${DB_CONNECTION_IDLE_TIMEOUT:600000}
        minimum-idle: ${DB_CONNECTION_MINIMUM_IDLE:30}
        maximum-pool-size: ${DB_CONNECTION_MAX_POOL_SIZE:30}

[sergiyvamz]

Hi! I recall I saw similar behaviour with MySQL Connector driver. Could you run a similar test without AWS Wrapper driver (i.e. by directly using PG driver)? It would help to narrow down the issue. Thank you!