Database connector: Oracle stored procedure calls silently dropping connections after 8 minutes post CloudHub 2.0 migration.

After migrating to CloudHub 2.0, an issue has been observed where stored procedure calls to a backend Oracle database—invoked through the Database connector and JDBC driver—unexpectedly terminate after approximately eight minutes. Notably, these terminations occur without any corresponding error logs, making troubleshooting particularly challenging.

The root cause of this behavior lies at the network layer and is specific to the CloudHub 2.0 infrastructure. During long-running stored procedure executions, the TCP connection can remain idle for extended periods. In such cases, an intermediate network component—such as a load balancer, NAT gateway, or firewall—may silently drop the connection after around eight minutes of inactivity. This results in the abrupt termination of the database call without any application-level error reporting.