psycopg2 connection remains in hung state during RDS failover

[roshanpathak500]

I am trying to connect to postgres database (hosted on AWS RDS server) from python application .

The issue is whenever there is a RDS reboot with failover , the code gets stuck for an hour without throwing "SSL SYSCALL error: Operation timed out" exception . Not just this, the hung state blocks the main process on the EKS pod in which python application is running in turn not allowing any requests to be honored .
On the Contrary, when I simply shutdown the RDS (instead of failover) , the process immediately fails and shows the exception allowing me to do further operations .

from psycopg2.pool import SimpleConnectionPool

def get_psycopg2_conn():
        try:
            connection_pool = SimpleConnectionPool(
                minconn=5,
                maxconn=30,
                keepalives=1,
                keepalives_idle=30,
                keepalives_interval=10,
                keepalives_count=5,
                host=“myhost”,
                database=“myappdb”,
                user=“dummyuser”,
                password=“dummypassword”
            )
            return connection_pool.getconn()
        except:
            return None

Removing any pool strategy also does not work .

 conn = psycopg2.connect(
                host="myhost"
                database="myappdb",
                user="dummyuser",
                password=""dummypassword"            
                )

What can be done to ensure the execution does not go into hung state rather throws exception right away ?

[dvarrazzo]

We cannot support RDS for free: we cannot run any test for free, as it is a paid-for service.

If you want to purchase consultancy time feel free to reach out privately.