A traditional CI/CD pipeline pushes changes out to a cluster — GitOps flips that around: a controller running inside the cluster pulls its desired state from Git and reconciles reality to match it. This module covers that model, starting with Git as the single source of truth and the declarative deployments and reconciliation loops that make it work, using tools like ArgoCD and Flux.
It also covers artifact repositories like Nexus, Artifactory, and container registries for storing what your pipelines build, semantic versioning and release management, safe rollbacks and progressive delivery, running GitOps across multiple environments, and the access control considerations that come with giving Git that much authority over your infrastructure.
Watch the lessons in order, then continue on to Secret Management.
The core idea behind GitOps — declaring desired state in Git and letting a controller pull and apply it — and how it differs from a push-based pipeline.
Setting up a GitOps controller to watch a repository and automatically sync a Kubernetes cluster to match it.
How a reconciliation loop continuously compares actual state to desired state and corrects drift automatically, without manual intervention.
Storing and versioning the build outputs your pipelines produce — packages, binaries, and container images — in a dedicated repository.
Using major.minor.patch versioning to communicate the impact of a change, and managing releases in a way that stays predictable for consumers.
Reverting a bad deployment quickly by rolling back to a previous Git commit, and rolling out changes gradually to limit risk.
Structuring repositories and branches to manage dev, staging, and production environments cleanly within a GitOps model.
Protecting the repository that now controls your infrastructure — branch protection, signed commits, and limiting who can merge to it.