A simple example showing multi-stage deployments using GitHub Actions.

The testing and deployment logic present in script/test and script/deploy are merely illustrative and need to be modified to do anything meaningful.

This example uses the master branch to deploy staging and a production branch to deploy production. These conventions can be changed to use any refs you'd like; just change the on settings in .github/workflows/deploy-*.yml.

Various repository settings (specifically those related to branch protection and status check enforcement) should be set as well; some suggestions are made below.

Create a topic branch:

$ git checkout -b change-abc

Make your changes...

Then, commit and push:

$ git commit -m "Changed ABC"
$ git push --set-upstream origin change-abc

On GitHub, create a pull request.

⚙️ The CI workflow will run. (Tip: In your repo, set Require status checks to pass before merging for ci on your master branch by visiting Settings ➡️ Branches ➡️ Add rule to prevent non-passing PRs from being merged.)

When CI passes (and your code is ready, reviewed, etc), merge the PR.

⚙️ The Deploy to Staging workflow will run. Note that this workflow will also run the tests to ensure they pass post-merge before deploying.

Once deployed to staging, after any and all manual testing is complete, promote staging to production by merging your changes from master into the production branch. One way to do this is by a specific deployment branch:

# While on the 'master' branch; just pushing this to an alias
$ git push origin master:change-abc-deploy

Then submitting a pull request to merge change-abc-deploy into production.

⚙️ The CI workflow will run. (Tip: In your repo, set Require status checks to pass before merging for ci on your production branch by visiting Settings ➡️ Branches ➡️ Add rule to prevent non-passing PRs from being merged.)

When 100% ready, merge the pull request.

⚙️ The Deploy to Production workflow will run. Note that this workflow will also run the tests to ensure they pass post-merge before deploying.