Split up Deploy workflow into multiple jobs

[GaryGSC]

I added a hadolint job under the assumption that #705 would be accepted.

Here, we take several steps that were part of the one and only job, and split them out into separate jobs so that those steps can be run in parallel. This parallelization should reduce deployment times.

This was a simple first attempt, but there is some further work that we could do:

• Use reusable workflows to DRY up some code shared between CI and deployments
• Move Docker build/push into a separate job
  • Potentially complicated because ideally we'd have a single build with multiple pushes for multiple environments
• Move TF plan analysis to a separate job
  • Potentially complicated because we'd like to re-use the plan, which might get weird with the way terraform-setup wraps outputs

Closes #541
Closes #677

[joshgubler]

Faster terraform fmt!

[GaryGSC]

On the plus side, we no longer test/audit/lint multiple times if we're deploying to multiple environments!