Assumptions

• ArgoCD is used to deploy artifacts (apps, ingress controller and ingress resources, etc.) onto dev, qa and prod clusters,
• ArgoCD resources are deployed onto argocd cluster
• two AWS accounts to separate production env (not obligatory),
• docker images are built by any other tool / pipeline defined outside of this repo,
• k8s clusters are already created and kubectl access to all of them is ready,

Requirements

• kubectl
• awscli v2

Installation

ArgoCD CLI

make install/cli

ArgoCD k8s resources

To deploy all ArgoCD related resources, call:

kubectl config set-context argocd
make install/argocd
make update/apps

Configuration

To get access to ArgoCD UI at http://localhost:8080, call:

kubectl --context argocd port-forward svc/argocd-server 8080:80

Initial password:

kubectl --context argocd -n argocd get secret argocd-initial-admin-secret -o jsonpath="{.data.password}" | base64 -d; echo

Login CLI:

argocd login localhost:8080

Add more clusters:

Notice: we assume your ~/.kube/config file contains dev and prod contexts already

argocd cluster add dev
argocd cluster add prod

The above commands install a ServiceAccount (argocd-manager), into the kube-system namespace of that kubectl context, and binds the service account to an admin-level ClusterRole. Argo CD uses this service account token to perform its management tasks (i.e. deploy/monitoring).

Deployment

You're now good to go with the deployments. All ArgoCD apps are defined as helm charts in ./apps/ folder and are managed as ApplicationSet - see applicationset.yaml for more details.

If you wish to add another deployable artifact, add a new folder with a helm chart inside.

To deploy make changes in main branch within ./apps/ folder.