Force botocore to connect using ipv6 first when using dualstack

[robd003]

Describe the bug

When using an EC2 instance with a private ipv4 network and a public ipv6 network botocore will timeout trying to connect to the ipv4 address if use_dualstack=True

If there was a way to prefer IPv6 when dualstack is enabled that would be great.

Regression Issue

• Select this option if this issue appears to be a regression.

Expected Behavior

Attempt to connect with IPv6 first with a fallback to IPv4 when use_dualstack=True

Current Behavior

Program hangs until botocore endpoint connection timeout happens because ipv4 isn't routable

Reproduction Steps

Create ec2 instance with private ipv4 and public ipv6
Attempt to connect to any dualstack AWS service

Possible Solution

Terrible workaround

Run this before calling botocore:

import socket

def patched_getaddrinfo(*args, **kwargs):
    return socket.getaddrinfo(*args, family=socket.AF_INET6, **kwargs)

socket.getaddrinfo = patched_getaddrinfo

Additional Information/Context

No response

SDK version used

1.36.4

Environment details (OS name and version, etc.)

Linux in EC2 (any distribution, all will have this same problem), Python 3.13