HIKARICP：为什么连接泄漏和不泄漏导致新DB连接“连接尝试超时”？

Question

我正在运行一个Spring boot Java应用程序，使用默认的HikariCP作为数据源：

• Hikari版本-3.4.5
• 在AWS EKS中运行的Spring Boot version-2.4.5
• JDBI版本-3.9.1
• DB-AWS RDS Postgres

我的Spring应用程序变得没有响应，因为它失去了Hikari池中的所有连接。这并不是确定地发生，但是一旦它开始发生，唯一可能的恢复是通过重新启动。

06-23 08:14:33.242 [HikariPool-1 housekeeper] DEBUG HikariPool - HikariPool-1 - Pool stats (total=10, active=1, idle=9, waiting=0)
06-23 08:15:07.189 [HikariPool-1 housekeeper] DEBUG HikariPool - HikariPool-1 - Pool stats (total=10, active=1, idle=9, waiting=0)
06-23 08:15:07.190 [HikariPool-1 housekeeper] DEBUG HikariPool - HikariPool-1 - Fill pool skipped, pool is at sufficient level.
06-23 08:15:32.785 [HikariPool-1 connection closer] DEBUG PoolBase - HikariPool-1 - Closing connection org.postgresql.jdbc.PgConnection@58f866df: (connection was evicted)
06-23 08:15:41.278 [HikariPool-1 housekeeper] DEBUG HikariPool - HikariPool-1 - Pool stats (total=9, active=1, idle=8, waiting=0)
06-23 08:15:49.809 [HikariPool-1 connection adder] DEBUG PoolBase - HikariPool-1 - Failed to create/setup connection: Connection attempt timed out.
06-23 08:15:49.809 [HikariPool-1 connection adder] DEBUG HikariPool - HikariPool-1 - Cannot acquire connection from data source
org.postgresql.util.PSQLException: Connection attempt timed out.
        at org.postgresql.Driver$ConnectThread.getResult(Driver.java:422)
        at org.postgresql.Driver.connect(Driver.java:267)
        at com.zaxxer.hikari.util.DriverDataSource.getConnection(DriverDataSource.java:138)
        at com.zaxxer.hikari.pool.PoolBase.newConnection(PoolBase.java:358)
        at com.zaxxer.hikari.pool.PoolBase.newPoolEntry(PoolBase.java:206)
        at com.zaxxer.hikari.pool.HikariPool.createPoolEntry(HikariPool.java:477)
        at com.zaxxer.hikari.pool.HikariPool.access$100(HikariPool.java:71)
        at com.zaxxer.hikari.pool.HikariPool$PoolEntryCreator.call(HikariPool.java:725)
        at com.zaxxer.hikari.pool.HikariPool$PoolEntryCreator.call(HikariPool.java:711)
        at java.base/java.util.concurrent.FutureTask.run(FutureTask.java:264)
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
        at java.base/java.lang.Thread.run(Thread.java:829)
06-23 08:15:49.810 [HikariPool-1 connection adder] DEBUG HikariPool - HikariPool-1 - Connection add failed, sleeping with backoff: 250ms
06-23 08:15:49.816 [HikariPool-1 connection closer] DEBUG PoolBase - HikariPool-1 - Closing connection org.postgresql.jdbc.PgConnection@7303422e: (connection was evicted)

(After more of these)
06-23 10:23:21.731 [ForkJoinPool-1-worker-7] DEBUG HikariPool - HikariPool-1 - Add connection elided, waiting 2, queue 10
06-23 10:23:55.883 [ForkJoinPool-1-worker-7] DEBUG HikariPool - HikariPool-1 - Timeout failure stats (total=0, active=0, idle=0, waiting=0)
06-23 10:23:55.883 [pool-3-thread-4] DEBUG HikariPool - HikariPool-1 - Timeout failure stats (total=0, active=0, idle=0, waiting=0)
06-23 10:23:55.883 [HikariPool-1 housekeeper] DEBUG HikariPool - HikariPool-1 - Pool stats (total=0, active=0, idle=0, waiting=0)
06-23 10:23:55.883 [HikariPool-1 housekeeper] DEBUG HikariPool - HikariPool-1 - Fill pool skipped, pool is at sufficient level.

06-23 08:07:04.506 [HikariPool-1 housekeeper] WARN  ProxyLeakTask - Connection leak detection triggered for org.postgresql.jdbc.PgConnection@58f866df on thread ForkJoinPool-1-worker-1, stack trace follows
java.lang.Exception: Apparent connection leak detected
        at com.zaxxer.hikari.HikariDataSource.getConnection(HikariDataSource.java:128)
        at org.springframework.jdbc.datasource.DataSourceUtils.fetchConnection(DataSourceUtils.java:158)
        at org.springframework.jdbc.datasource.DataSourceUtils.doGetConnection(DataSourceUtils.java:116)
        at org.springframework.jdbc.datasource.TransactionAwareDataSourceProxy$TransactionAwareInvocationHandler.invoke(TransactionAwareDataSourceProxy.java:223)
        at com.sun.proxy.$Proxy119.getAutoCommit(Unknown Source)
        at org.jdbi.v3.core.transaction.LocalTransactionHandler.isInTransaction(LocalTransactionHandler.java:118)
        at org.jdbi.v3.core.Handle.(Handle.java:76)
        at org.jdbi.v3.core.Jdbi.open(Jdbi.java:309)
        (Application specific stacktrace)
...

06-23 08:07:04.527 [ForkJoinPool-1-worker-1] INFO  ProxyLeakTask - Previously reported leaked connection org.postgresql.jdbc.PgConnection@58f866df on thread ForkJoinPool-1-worker-1 was returned to the pool (unleaked)

1. 连接泄漏是问题的根源吗？因为它们是未泄露的，Hikari难道不能收回连接并创建新的连接吗？有办法确认根本原因吗？
2. 泄漏并不经常发生，但当它们发生时，它们会在几个请求中开始发生，直到应用程序变得没有响应。是否有办法找出请求开始超过20秒的原因？
3. 哪些配置更改有助于进行更多的诊断并进一步解决问题？

谢了。

Answer 1

如果你不关闭连接，hikari不能知道你需要/想要关闭它

您必须找到在代码中没有正确关闭连接的地方

您可以通过设置较低的maxLifetime来减少泄漏，以便更快地将连接返回到池