Automatic Deployment Workflow

Example of a fully automated multi-environment deployment pipeline using GitHub Actions, Terraform, and Azure. Written about in detail in the companion blog post: Automating deployments to production and staging with GitHub Workflows.

How it works

Pushes to development deploy to staging; pushes to main deploy to production. The workflow only deploys what actually changed:

1. Check changes – a composite action diffs the last commit against backend/, frontend/, and infrastructure/ to decide which jobs run.
2. Set environment – maps the branch (or PR target) to staging or production.
3. Terraform plan – runs when infrastructure changed; the plan is published to the job summary and posted as a PR comment.
4. Terraform apply – applies the uploaded plan on main/development, using a per-environment state backend.
5. Deploy backend – builds the .NET Azure Function and publishes it to the environment's Function App.
6. Deploy frontend – builds the Svelte app and deploys it to Azure Static Web Apps.

Backend and frontend deploys also run standalone when only their directory changed, so an app-only commit never waits on Terraform.

Repository layout

Directory	Contents
.github/	Deployment workflow + composite actions (change detection, environment selection)
backend/	.NET Azure Function app
frontend/	Svelte + Vite app served from Azure Static Web Apps
infrastructure/	Terraform for all Azure resources, parameterized by environment

Reusing this

Fork it, create staging/production GitHub environments, and set the Azure credential secrets referenced in deployment.yml (AZURE_CLIENT_ID, AZURE_SUBSCRIPTION_ID, AZURE_TENANT_ID, AZURE_CLIENT_SECRET, AZURE_CREDENTIALS, PAT, plus the Static Web Apps API keys). The same pattern works for any stack — swap the build steps and keep the change-detection/environment plumbing.