Part 1: Presentation and discussion: what problem do continuous integration and continuous deployment solve?

Deploying software directly to production sounds like a crazy idea. However, more often than not, this should be the default. Even in organizations where a CICD pipeline is established, there are still multiple "environments": the production environment, where the actual business is happening, the staging and maybe even development environments, where supposedly some QA is taking place, and, of course, there is always "it works on my laptop". In such organizations despite, having a CI/CD pipeline, software is eventually deployed by hand.

Part 2: Should we prefer deploying straight to production and how? Presentation and discussion:

We discuss why a pipeline with correct checks and balances should not be triggered manually, but automatically instead. We examine different strategies for building such pipeline and software integration tests. Finally, we discuss different continuous deployment strategies and examine real life cases.

Part 3: Practice building a pipeline using gitlab

Deploy a small web application using gitlab pipeline. We set up a small scale pipeline that will "build" the sources of our application in a form of a Docker container and then deploy the application to a live Kubernetes cluster.

Part 4: Practice deployment strategies

Exercise deploying different versions and of the application to the cluster using different deployment strategies and various mechanism. Once, the pipeline is set up, we can tweak it and use Kubernetes native tools to practice different deployment strategies. Once we trust Kubernetes native tools, we examine how we can do the same using Ingress controllers.