Kubernetes GitOps Repo

This is an example on how I would structure a 1:1 (repo-to-single cluster) git repo for a Kubernetes cluster. This is based on the OpenShift repo I created.

This example assumes (as I mentioned in the 1:1 part above) that it's a single repo for a single cluster. However, this can be modified (quite easily) for poly/mono repos or for multiple clusters. This is meant as a good starting point and not what your final repo will look like.

This is based on Argo CD but the same principals can be applied to Flux.

Structure

Below is an explanation on how this repo is laid out. You'll notice that I use Kustomize heavily. I do this since I follow the DRY principal when it comes to YAML files.

cluster-XXXX/ # 1
├── bootstrap # 2
│   ├── base
│   │   ├── argocd-ns.yaml
│   │   └── kustomization.yaml
│   └── overlays
│       └── default
│           └── kustomization.yaml
├── components # 3
│   ├── applicationsets
│   │   ├── core-components-appset.yaml
│   │   ├── kustomization.yaml
│   │   └── tenants-appset.yaml
│   └── argocdproj
│       ├── kustomization.yaml
│       └── test-project.yaml
├── core # 4
│   ├── gitops-controller
│   │   └── kustomization.yaml
│   └── sample-admin-config
│       ├── kustomization.yaml
│       └── sample-admin-config.yaml
└── tenants # 5
    ├── bgd-blue
    │   ├── bgd-deployment.yaml
    │   └── kustomization.yaml
    └── myapp
        ├── kustomization.yaml
        ├── myapp-deployment.yaml
        ├── myapp-ns.yaml
        └── myapp-service.yaml

#	Directory Name	Description
1.	cluster-XXXX	This is the cluster name. This name should be unique to the specific cluster you're targeting. If you're using CAPI, this should be the name of your cluster, the output of kubectl get cluster
2.	bootstrap	This is where bootstrapping specifc configurations are stored. These are items that get the cluster/automation started. They are usually install manifests. base is where are the "common" YAML would live and overlays are configurations specific to the cluster. The kustomization.yaml file in default has cluster-XXXX/components/applicationsets/ and cluster-XXXX/components/argocdproj/ as a part of it's bases config.
3.	components	This is where specific components for the GitOps Controller lives (in this case Argo CD). applicationsets is where all the ApplicationSets YAMLs live and argocdproj is where the ArgoAppProject YAMLs live. Other things that can live here are RBAC, Git repo, and other Argo CD specific configurations (each in their repsective directories).
4.	core	This is where YAML for the core functionality of the cluster live. Here is where the Kubernetes administrator will put things that is necissary for the functionality of the cluster (like cluster configs or cluster workloads). Under gitops-controller is where you are using Argo CD to manage itself. The kustomization.yaml file uses cluster-XXXX/bootstrap/overlays/default in it's bases configuration. This core directory gets deployed as an applicationset which can be found under cluster-XXXX/components/applicationsets/core-components-appset.yaml. To add a new "core functionality" workoad, one needs to add a directory with some yaml in the core directory. See the sample-admin-config directory as an example.
5.	tenants	This is where the workloads for this cluster live. Similar to core, the tenants directory gets loaded as part of an ApplicationSet that is under cluster-XXXX/components/applicationsets/tenants-appset.yaml. This is where Devlopers/Release Engineers do the work. They just need to commit a directory with some YAML and the applicationset takes care of creating the workload. Note that bgd-blue/kustomization.yaml file points to another Git repo. This is to show that you can host your YAML in one repo, or many repos.

Testing

To see this in action, first get yourself a cluster (using kind as an example)

kind create cluster

Then, just apply this repo.

until kubectl apply -k https://github.com/christianh814/example-kubernetes-go-repo/cluster-XXXX/bootstrap/overlays/default; do sleep 3; done

This should give you 4 applications

$ kubectl get applications -n argocd
NAME                          SYNC STATUS   HEALTH STATUS
bgd-blue                      Synced        Healthy
sample-admin-workload         Synced        Healthy
myapp                         Synced        Healthy
gitops-controller             Synced        Healthy

Backed by 2 applicationsets

$ kubectl get appsets -n argocd
NAME      AGE
cluster   110s
tenants   110s

To see the Argo CD UI, you'll first need the password

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

Then port-forward to see it in your browser (using admin as the username).

kubectl -n argocd port-forward service/argocd-server 8080:443

Enjoy

Fork and Enjoy!

⚠️ Feel free to fork this and play around with it, but remember if you'll have to change the applicationsets configuration and where the bases are pointing to if you're chanigng names of things