Access to VPC inside deployment docker runtime

Allow AWS Amplify deployment process to access existing default VPC or provide an option as AWS CodeBuild project has here:

https://aws.amazon.com/blogs/devops/access-resources-in-a-vpc-from-aws-codebuild-builds/

We tried to implement sequelize/typeorm database migrations inside AWS Amplify deployment pipeline, but failed to do so, because it's isolated from the existing VPC...

This would be great to have that there.

Start with the linked AWS CodeBuild VPC example and the AWS Amplify deployment architecture described in the issue. Determine where deployment runtime networking is configured and how an opt-in VPC setting could be exposed; done means migrations or other deployment steps can reach resources in an existing VPC without breaking deployments that do not use it.

Written by the indexing model from the issue text.