add HTTPCallbackAddress, needed when running within a kubernetes cluster

[lknite]

Describe the change you are making here!

When running within a kubernetes workflow/pipelines, the http server used by the vm to communicate back completion cannot be reached directly. Instead, with kubernetes you 'expose' a pod via a LoadBalancer ip or ingress. The exposed ip is available but must be passed in somehow in order for it to be used.

This pull request implements the use of a variable HTTPCallbackAddress, which if provided, is passed along to the vm to be used upon completion when calling back to the http server.

A pull request has been submitted to the packer-plugin-sdk to add the needed variable: https://github.com/hashicorp/packer-plugin-sdk/pull/268

Please include tests. Check out existent tests in the code for example.

I am open to what tests would be useful to add. The code change is very small, and it might be fair to say if there is a test for HTTPAddress, and it is passing, that might be good enough to verify this implementation as well.

If your PR resolves any open issue(s), please indicate them like this so they will be closed when your PR is merged:

Closes: #304