Deconstructing blue/green deployments in Kubernetes

Matthew Casperson's blog post explores how to perform blue/green deployments in Kubernetes using Octopus, beyond its native recreate and rolling deployment methods. The post details the process of implementing blue/green deployments manually, which allows for greater flexibility including halting deployments for testing, orchestrating multiple deployments, and using feature branches. Casperson demonstrates the procedure by deploying a mock feature branch with the httpd docker image, utilizing Kubernetes deployment resources and service configurations. He discusses creating variables to manage feature branches, setting up temporary and persistent services for testing and traffic redirection, and performing health checks with curl. The post concludes by explaining how to clean up old resources and suggests that the basic framework can be customized with additional steps for more complex deployment workflows.