Why ArgoCD?
With Kubernetes it’s easy to end up with a cluster nobody can fully explain: someone ran kubectl apply by hand, someone else patched a value at 2 a.m. ArgoCD fixes that by making Git the single source of truth. The desired state lives in a repository, ArgoCD keeps the cluster in sync with it and shows clearly when the two drift apart.
The result is that deployments become pull requests. They’re reviewed, logged and easy to roll back by reverting a commit.
How I Use It
- Exlink: I set up ArgoCD as part of the full migration to Kubernetes, so the platform’s services deploy from Git instead of setup scripts run by hand
- Luigi’s Box: GitHub Actions builds and tests, ArgoCD syncs the production clusters on AWS. I used this flow daily to deploy and configure internal services
- Drift detection: if someone changes the cluster outside Git, ArgoCD shows it immediately
- Rollbacks: reverting a commit is the rollback, no special procedure needed
Status
Active: the delivery layer I pair with every Kubernetes cluster I work on.