Allow AWS bucket name to be provided as an environment variable

Open timrogers opened this issue 2 years ago • 1 comments

For certain migration paths, a blob storage provider must be configured in the CLI for storing the migration archive and transferring it to GitHub.

When using AWS, all of the required blob storage connection settings can be provided using environment variables or CLI arguments, except the S3 bucket name which can only be provided using --aws-bucket-name.

Conversations with customers suggest that people are finding this confusing - "why can I specify everything AWS-related as an environment variable except the bucket name?". We should add the ability for this also be be specified in the environment.

Todo

[ ] ...

Dependencies

Mar 28 '23 17:03 timrogers

When we do this one, we'll have to consider what we do in generate-script should we require the user to pass it to generate-script then include it in the script as --aws-bucket-name. Or should we assume the user running the script will set it as an env var.

Mar 31 '23 16:03 dylan-smith