Argoproj - Application Delivery Platform

Argoproj is a set of loosely coupled tools that aims to unleash to power of Kubernetes:

• Argo Workflows - container-native workflow engine for orchestrating parallel jobs on Kubernetes.
• Argo Events - The event-driven workflow automation framework
• Argo CD - Declarative continuous delivery for Kubernetes
• Argo Rollouts - Progressive delivery for Kubernetes

Each Argo sub-project is focused on a separate use-case and can be used independently. But together, Argo projects complement each other and form a powerful application delivery platform.

Demo

This repository demonstrates how Argo allows gluing disconnected open-source projects into a complex system that solves real-world use cases, with zero code written.

We are going to build a web application consisting of multiple microservices as well as a background batch processing system that leverages machine learning. The application allows users to upload an image, stores the photo in an S3 compatible storage and produces a new image with highlighted faces on it using a pre-trained ML model.

Components

Instead of building an application from scratch, we are going to leverage existing open-source projects and the Argo projects to make them work together for profit:

• minio - Kubernetes native S3 compatible storage. Minio is going to store user uploaded images and ML workflows outputs.
• filestash - A modern web client for the storage of your choice. Filestash allows user to upload an image and view processing results.
• facedetect - a simple face detector for batch processing. The facedetect encapsulates the magic and produces an image with detected faces.
• argo-workflows - workflow engine for Kubernetes. Argo Workflows orchestrates steps requires to access uploaded image, process the image and store processing results.
• argo-events - event-driven automation framework for Kubernetes. Argo Events are watching for new images in the S3 storage and triggers the batch processing workflow.
• argo-cd - GitOps continuous delivery tool for Kubernetes. Argo CD manages components listed above it the Kubernetes cluster and encapsulates both day one and day two operations.

Let's do it

1. First of all we need a Kubernetes cluster. Nothing fancy is required here. Use your GKE, EKS cluster or just run minikube cluster on your laptop.
2. Install Argo CD. Follow the getting started instructions in Argo CD documentation.

TLDR:

kubectl create namespace argocd
kubectl apply -n argocd -f https://raw.githubusercontent.com/argoproj/argo-cd/stable/manifests/install.yaml

3. Deploy the demo scenario using Argo CD, which in turn will spawn all required components in your K8s cluster:

kubectl apply -f https://raw.githubusercontent.com/alexmt/argo-combined-demo/master/demo.yaml -n argocd

4. Check the components' status in Argo CD user interface.

Try it

1. Navigate to the external IP address of demo-filestash service on port 9001 and upload an image using the filestash web user interface. Note: image file must have ".jpg" extension.
2. Use argo workflows user interface to observe background processing process. Workflows user interface is available via external IP address of argo-server service on port 2746.
3. See image processing results in filestash interface.