GitOps
Definition
A deployment approach where infrastructure and application changes are managed through Git version control, promoting collaboration and efficiency.
Use Cases
- Weaveworks: Operating Kubernetes clusters with a Git-driven deployment and configuration workflow — Weaveworks popularized the GitOps approach and built Flux, a GitOps controller that watches Git repositories and applies declared Kubernetes state to clusters, enabling automated reconciliation and auditability. (Teams can standardize deployments, improve traceability of changes, and reduce manual cluster drift by making Git the source of truth.)
- Intuit: Scaling Kubernetes application delivery with standardized, auditable deployment workflows — Intuit has publicly discussed adopting Kubernetes and modern delivery practices, including Git-based workflows and automation, to manage changes consistently across environments. (Improved consistency and repeatability of deployments and better governance through version-controlled change history.)
- CERN: Managing large-scale Kubernetes infrastructure and application deployments — CERN has shared Kubernetes operations practices publicly, including Git-based configuration management and automation patterns aligned with GitOps-style workflows for cluster and application state. (More reliable operations through reproducible configuration, easier rollbacks, and clearer change auditing.)
Frequently Asked Questions
- What’s the difference between GitOps and CI/CD?
- CI/CD is the automation pipeline that builds, tests, and delivers software (for example, building a container image and running tests). GitOps is an operating model where Git is the source of truth for what should be running, and an automated controller continuously applies and reconciles the live environment to match what’s in Git. In practice, CI/CD often produces artifacts (like images), while GitOps applies the desired state (like Kubernetes manifests or Helm values) from Git.
- When should I use GitOps?
- Use GitOps when you want consistent, auditable deployments and configuration management—especially for Kubernetes. It’s a good fit if you have multiple environments (dev/stage/prod), need strong change control (pull requests, approvals, history), want easy rollbacks, or frequently see “configuration drift” from manual changes. It may be less useful for very small projects where the overhead of controllers, repo structure, and policies outweighs the benefits.
- How much does GitOps cost?
- GitOps itself doesn’t have a fixed cost—it depends on the tools and infrastructure you choose. Common cost factors include: Git hosting (often free/paid tiers), CI runners/build minutes, the compute resources for the GitOps controller (e.g., Argo CD/Flux running in your cluster), and any managed add-ons (observability, policy, secrets management). If you run everything yourself on Kubernetes, the main cost is the cluster resources and operational time; commercial GitOps platforms add licensing or subscription fees.
Category: emerging
Difficulty: advanced
Related Terms
See Also