Setting up an ArgoCD instance

This how-to explains how to create a new ArgoCD instance on your cluster, leveraging the ArgoCD operator which comes preinstalled on APPUiO Managed OpenShift clusters.

Requirements

You will need admin access to the cluster on which you want to deploy ArgoCD.

Furthermore, this how-to assumes you have the following tools installed:

oc

You can download the OpenShift command directly from OpenShift Web Console. Open the help menu (marked as a question mark at the top right) and select the "Command line tools" entry.

Cluster Scope vs. Namespace Scope

By default, an ArgoCD instance will be able to manage all resources within the same namespace it’s deployed in, making it namespace scoped. This allows for easy setup of small-scope GitOps projects.

However, ArgoCD is also capable of managing resources across multiple namespaces and cluster-scoped resources. To accomplish this, an ArgoCD instance needs to be registered as cluster scoped in the VSHN-managed ArgoCD operator.

Requesting a Cluster-Scoped ArgoCD instance

If you need to manage resources across multiple namespaces with the same ArgoCD instance, or you need to manage cluster-scoped resources, you will need to request a cluster-scoped instance.

Create a Service Desk Ticket with the following information: * State that you need a cluster-scoped ArgoCD instance * Include the namespace name in which your ArgoCD instance will be deployed

VSHN will register the provided namespace to allow for a cluster-scoped ArgoCD instance. Once that’s complete, you can set up your ArgoCD instance as described below; it will automatically be cluster scoped.

Setting up an ArgoCD instance

Thanks to the ArgoCD operator, new ArgoCD instances can easily be set up by deploying an argocd manifest.

  1. Create the namespace for ArgoCD, or select it if it already exists

    oc new-project my-argocd-namespace
    oc project my-argocd-namespace

  2. Create ArgoCD manifest

    Manifest for ArgoCD instance argocd.yaml
    apiVersion: argoproj.io/v1beta1
kind: ArgoCD
metadata:
  name: my-argocd
spec:
  applicationInstanceLabelKey: mycompany.com/appname (1)
    1 The applicationInstanceLabelKey is the name of a label, which your ArgoCD instance will use to keep track of the resources it manages. You must configure this label key and use a unique name for it.
    The default label key is already in use by the VSHN-managed platform ArgoCD instance. Using the same label key in your instance might lead to deletion of important resources.

  3. Customize the ArgoCD manifest to suit your needs, as per the upstream reference documentation.

  4. Deploy the ArgoCD manifest

    oc apply -f argocd.yaml --as cluster-admin

  5. Verify the ArgoCD deployment

    oc get pods

    You should see a number of pods starting up for your ArgoCD instance, including the ArgoCD server, application controller, and repo server. The exact constellation of pods depends on your configuration.

Accessing the ArgoCD UI

The ArgoCD UI is available at the ArgoCD server service.

  1. Retrieve the admin password from the initial password secret (substituting the appropriate name for your ArgoCD instance):

    oc get secret my-argocd-cluster -ojsonpath='{.data.admin\.password}' | base64 -d

  2. Port-forward the server service (substituting the appropriate name for your ArgoCD instance):

    oc port-forward service/my-argocd-server 8080:80

  3. Access ArgoCD using your browser at localhost:8080

  4. Log in with user admin and the password retrieved earlier