Frequently Asked Questions
1. What is GitOps?
GitOps is a way to do operations, by using Git as a single source of truth, and updating the state of the operating configuration automatically, based on a Git repository.
2. How does GitOps differ from Infrastructure as Code?
GitOps builds on top of Infrastructure as Code, providing application level concerns, as well as an operations model.
3. Can I use a CI server to orchestrate convergence in the cluster?
You could apply updates to the cluster from the CI server, but it won’t continuously deploy the changes to the cluster, which means that drift won’t be detected and corrected.
4. Should I abandon my CI tool?
No, you'll want CI to validate the changes that GitOps is applying.
5. Why choose Git and not a Configuration Database instead? / Why is git the source of truth?
Git has strong auditability, and it fits naturally into a developer's flow.
6. How do you keep my tokens secret in the Git repository?
We are going with Sealed Secrets because of it's low-maintenance, and because it requires little investment to get going.
7. How do I get started?
Add some resources to a directory, and git commit and push, then ask ArgoCD to deploy the repository, change your resource, git commit and push, and the change should be deployed automatically.
8. How are OpenShift pipelines (Tekton) used?
They are used in the default setup to drive the CI from pushes to your application code repository.
9. How is GitOps different from DevOps?
GitOps is a subset of DevOps, specifically focussed on deploying the application (and infrastructure) through a Git flow-like process.
10. How could small teams benefit from GitOps?
GitOps is about speeding up application feedback loops, with more automation, it frees up developers to work on the product features that customers love.