- Luca Cavallin
GitOps is known as an operational framework that takes the best practices of DevOps, such as CI/CD, version control and collaboration and applies them to infrastructure automation. GitOps is a set of code-based practices that use Git, an open-source version control system, to manage infrastructure and application configurations. Git is the single source of truth, and pull requests are used to verify and automatically manage/deploy infrastructure changes. Just like developers use application source code, operation teams that practice GitOps use infrastructure as code (configurations files stored as code) to generate the same infrastructure environment during each deployment. In short, GitOps uses similar processes and tools used in software development to manage infrastructure, ensuring the automation that teams need.
How does GitOps work?
Since GitOps involves Git as a version control system, it can be considered an advancement in Infrastructure as Code (IaC). In GitOps, changes are triggered via pull requests that change the state in the Git repository.
A GitOps workflow for updating or creating new feature is as follow:
Initiate a pull request for the new feature in Git
Review code and merge it to Git
Git will automatically trigger CI and build pipelines, run tests, apply infrastructure changes and create -when in a container-based environment- a new image which is then uploaded to the registry
Specifically for applications, deployment tools can be used to automatically update the running version on, for example, a Kubernetes cluster or another serverless product
The workflows of GitOps are meant to improve productivity and speed of development and deployment, along with ensuring that systems remain stable and reliable!
Pros and Cons of GitOps
Following are the major pros and cons associated with GitOps:
It enhances the DevOps team's productivity, as they can instantly deploy new configurations of infrastructures. If the changes aren't performing as required, the Git history lets the team easily revert to a stable state
It ensures faster deployment, as the team does not have to switch between tools for deploying the application (application and infrastructure changes can be deployed using the same tools)
It brings end-to-end standardization to the workflow
It can scale to hundreds of Kubernetes clusters
Since Git is used for storing the complete information of the deployed infrastructure, changes can easily be tracked and also it encourages a culture of sharing knowledge in teams
With automation in infrastructure definition and testing, lots of manual work go away. The team becomes more productive, while downtime gets reduced due to rollback/revert capability. GitOps can lower costs significantly!
GitOps encourages collaboration, but it also sometimes becomes a time-consuming and tedious job. For example, the approval process in GitOps involves many stages, such as creating a merge request, approving the changes, and deploying the changes. Engineers that are used to doing manual and fast changes might feel the whole process is time-consuming.
Collaborative culture also requires discipline from all the members to ensure commitment to the process. Moreover, teams are required to write and formalize everything so that GitOps can work perfectly.
GitOps is a powerful framework for managing modern infrastructure with a focus on the developer experience. It enables infrastructure management from the same version control system involved in application development, empowering teams to have a central collaborative environment with improved system reliability and stability.