If you use GitHub Actions to deploy in a non-continuous way (i.e. you trigger a deployment workflow manually) and you utilize
GitHub's Deployment Environments feature, you will likely end up that the ref
associated with the resulting deployment
object
in GitHub's data model points to the state of main
at the time your workflow ran, regardless of the ref that was actually deployed
(which is frequently a release tag).
This action looks for the most recent release which was created for the SHA which triggered the current workflow, mixes it's data with data that is input to the action, and creates a new deployment with the updated data. The old release is then marked inactive and subsequently deleted, preventing duplication of deployments.
The workflow below is built to allow manual triggering while providing a release tag and an environment name
name: Manually Deploy to Env
on:
workflow_dispatch:
inputs:
# you probably want to validate this input somewhere!
release_tag:
description: the tag of the release to deploy
type: string
required: true
environment:
description: the environment to deploy to
type: choice
options:
- production
required: true
jobs:
# Do your deployment as appropriate, setting the environment so that a deployment is created
do_deploy:
environment: ${{ inputs.environment }}
runs-on: ubuntu-latest
steps:
- run: echo "Do deployment of ${{inputs.release_tag}} to ${{ inputs.environment }}"
# run this action _in a dependent job_ so that it can find the newly created deployment from the previous job
clean_up_deploy:
runs-on: ubuntu-latest
needs: [do_deploy]
steps:
# be sure to replace with whatever version you'd like to pin to
- uses: kylebjordahl/[email protected]
with:
ref: ${{inputs.release_tag}}
description: Deployment of ${{inputs.release_tag}} to ${{ inputs.environment }}
environment: ${{ inputs.environment }}
token: ${{ secrets.GITHUB_TOKEN }}