devops/deployment-process

Create a workflow file triggered on push that runs install, test, build, and deploy steps in sequence on an Ubuntu runner. This Skill provides a ready-to-adapt GitHub Actions YAML example covering the full pipeline from code commit to production deployment.

Rolling deployment replaces instances gradually and suits zero-downtime releases. Blue-green deployment fits cases needing fast rollback via environment switching, while canary deployment works best for high-risk updates validated on small traffic first.

Blue-green deployment maintains two full environments and switches traffic between them, enabling instant rollback. Canary deployment routes a small portion of traffic to the new version first, reducing blast radius for risky updates.

Use test data in local and dev environments, a copy of production data in staging for realistic pre-release validation, and real data only in production. This separation prevents test pollution while keeping staging representative.

Choose canary deployment when the update carries higher risk and you want to validate behavior on a small traffic subset before full rollout. Rolling deployment is better for routine updates where gradual instance replacement is sufficient.