Reconsider image name/tag in light of managing different app environments

[rsyring]

One of the intentions of mu is that you can easily deploy multiple lambdas into the same AWS environment without conflict. This seems helpful, in particular, for being able to deploy/test different branches.

However, currently, we don't take the environment into account when creating images. So, the following sequence could result in accidentally deploying a dev branch to production:

1. Make changes on some-dev-branch, commit, provision/build/deploy with 'some-dev-branch' as the name
2. Switch to master / main, make a hot fix, deploy to prod (forgetting to build first)

The image sent to prod is now the last image built on the local system, which is the one for some-dev-branch.