Deploying Argo Workflows with Helm on Amazon EKS#

This guide explains how to deploy Argo Workflows on an Amazon EKS cluster. Argo Workflows is an open-source container-native workflow engine for orchestrating parallel jobs on Kubernetes.

Prerequisites#

Before deploying Argo Workflows, ensure you have the following prerequisites:

  • An Amazon EKS cluster. Refer to the kubernetes_deployment.rst in the project directory for instructions on deploying an EKS cluster with Terraform.

  • kubectl configured to interact with the deployed EKS cluster.

Local installation#

  1. Download argo cli from the official website <https://argo-workflows.readthedocs.io/en/latest/>

  2. Move the argo.exe to your directory of preference, here we will say C:\\cloud\\argo.

  3. Add said location to your PATH variables.

Deployment Steps#

Follow these steps to deploy Argo Workflows on the Amazon EKS cluster:

  1. Create Argo namespace:

    Create a namespace for Argo to run in:

    kubectl create namespace argo

  2. Install Argo Workflows:

    Update Helm repositories to ensure you have the latest information:

    kubectl apply -n argo -f https://github.com/argoproj/argo-workflows/releases/download/v3.5.5/install.yaml

  3. Access Argo UI:

    Once the installation is complete, you can access the Argo UI by port-forwarding to the Argo server service:

    kubectl -n argo port-forward service/argo-server 2746:2746

    Open your web browser and navigate to http://localhost:2746 to access the Argo UI.

Clean Up#

To uninstall Argo Workflows from the EKS cluster:

  1. Uninstall Argo Workflows:

    kubectl delete deployment argo -n argo

    This command removes the Argo Workflows deployment from the cluster.