Browse
GitOps
Using a Git repository as the single source of truth for infrastructure and deployment state.
What it is
GitOps uses a Git repository as the single source of truth for a system's desired infrastructure and application state — an automated agent continuously reconciles the live system to match what's declared in Git.
Key points
- Pull-based reconciliation: an in-cluster agent (e.g. Argo CD, Flux) watches the Git repo and applies changes, rather than an external CI pipeline pushing changes to the cluster — this is the key architectural difference from a traditional CD pipeline.
- Git as audit log: every change to infrastructure has a corresponding commit, giving a full history and an easy path to review (pull requests) and rollback (revert a commit).
- Drift detection: because the agent continuously reconciles, any manual out-of-band change to the live system gets automatically detected and reverted back to match Git — directly addressing configuration drift.
- GitOps is most associated with Kubernetes but the underlying idea — declarative desired state, versioned in Git, continuously reconciled — generalizes beyond it.
