I spoke with some of the other developers, and we will need to change a setting on our end to have our proxy end point support the http_proxy variables. We hadn’t done this yet, because SSH is needed and typically not allowed out in proxied environments, so the setting wouldn’t have enabled the feature. Since your team has peering with AWS, the SSH part should not be an issue for you. We are looking to add the setting in a near future release.
In the meantime, you can still use Rancher with AWS, you can provision an instance as you normally would through AWS console or AWS cli and run the “Custom command” on the host. You can find the custom command under the “Add Host” button, and instead of AWS, select “Custom”.
To automate the process, we have seen simple cloud-init scripts passed through user-data work well. Some others with full on configuration management registering the agent etc. Also, we have seen users create autoscaling groups that spin up nodes that auto-register with Rancher. Lots of options.
Please let me know if you have any questions, or run into issues. We will let you know when the setting has made it into an RC release.