How to Store ArgoCD Projects in YAML

ArgoCD Projects

This guide will explain the various components of ArgoCD Projects in YAML. In the spirit of GitOps, you may leverage deploying ArgoCD Projects from source code. This can be a powerful way to maintain Argo, and even consistently deploy to multiple clusters maintained by multiple Argo instances.

Understanding the ArgoCD Project

kind: AppProject
  name: hello-haze-project
  namespace: argocd
  description: AppProject for the Hello Haze UI
  - namespace: hello-haze
    server: https://kubernetes.default.svc
  - group: ""
    kind: ResourceQuota
  - group: ""
    kind: LimitRange
  - group: '*'
    kind: '*'
  - '**'

You may see some expected properties in this resource, but let me elaborate on these elements.

Line 2: This deploys an AppProject resource

Line 4-5: The project will be named hello-haze-project and deployed in the argocd namespace. In ArgoCD 2.5.0, Argo resources need to be in the argocd namespace. Future releases will likely ease this restriction.

Line 7: A simple description on the resource, later displayed in the Argo UI.

Lines 8-10: In this case, this Project only allows resources to deploy to the hello-haze namespace. Any attempts to deploy elsewhere will result in an error.

Lines 11-15: This block prohibits modifying the ResourceQuota and LimitRange within the namespace. This prevents any application from messing with resource restrictions place on the namespace.

Lines 16-18: This block allows all other resources to deploy to the destinations.

Lines 19-20: The sourceRepos are the Git repositories which this Project may read.

To learn more about configuring the Project, check out my post on Project configuration.