Loading...
Quick Definition
GitOps is an operational model that uses Git as the single source of truth for infrastructure and application configuration. Changes are made via pull requests and automatically reconciled to the live environment, bringing version control, auditability, and automation to operations.
GitOps is an operating model where the desired state of infrastructure and applications lives in a Git repository, and automated agents continuously make the real environment match it. To change production, you change a file and merge a pull request. The agent notices the difference and applies it.
The model builds on infrastructure as code and adds continuous reconciliation. If someone changes production by hand, the agent detects the drift and reverts it. Every change is reviewed, versioned, and reversible with a Git revert, which gives auditors a complete history for free.
Example. An engineer needs to scale a service from 4 to 8 replicas. Instead of clicking in a console, she edits one line in a YAML file and opens a pull request. A teammate approves, the merge lands, and the cluster updates itself within a minute.
GitOps also helps cost: configuration changes that raise spend become visible in review before they ship. See CI/CD best practices and the infrastructure as code primer for the foundations.
CI/CD pushes changes out through pipelines. GitOps adds a pull-based agent that constantly reconciles the environment against the repository.
Argo CD and Flux are the standard choices for Kubernetes. Terraform with automated plans serves a similar role for broader infrastructure.
It grew up there and fits it best, but the principle of Git as the source of truth applies to any declaratively managed system.