Azure DevOps — CI/CD for example-voting-app

This repository implements a CI/CD workflow for the multi micro-service "example-voting-app" (from DockerSamples) using Azure DevOps for CI, ArgoCD for continuous delivery, and Kubernetes for runtime deployment.

The goal is to demonstrate a full pipeline: build container images for each microservice, push them to a container registry from Azure DevOps pipelines, and deploy/manage the application in a Kubernetes cluster using ArgoCD.

This repo contains Azure DevOps pipeline YAMLs and helper scripts to automate manifest updates.

Repository layout

• pipelines/

  • results_pipeline.yml — CI pipeline for the results service
  • vote_pipeline.yml — CI pipeline for the vote service
  • worker_pipeline.yml — CI pipeline for the worker service

• scripts/

  • updatek8smanifest.sh — helper script to update image tags in Kubernetes manifests (used by CI/CD to pin images)

Note: The application source and original manifests are based on the upstream example at: https://github.com/dockersamples/example-voting-app/tree/main

High-level architecture

• Multiple microservices (vote, result, worker, redis, db, nginx) run as separate containers.
• CI (Azure DevOps): builds Docker images for each service, and pushes images to a container registry (Azure Container Registry, Docker Hub, or other).
• CD (ArgoCD): watches a Git repository containing Kubernetes manifests (this repo or a separate repo) and synchronizes the cluster state to the desired manifests. Image tags are updated by CI (using updatek8smanifest.sh) and committed to the manifests branch that ArgoCD watches.

CI pipelines (Azure DevOps)

Each YAML under pipelines/ corresponds to a service-specific CI pipeline. Typical responsibilities:

• Checkout source code
• Build the Docker image for the service
• Run unit/integration tests (if configured)
• Tag the image (e.g., $(Build.BuildId) or semantic tag)
• Push the image to a container registry
• Optionally update Kubernetes manifests with the new image tag and push those changes to the manifests branch for ArgoCD to pick up

Configure Azure DevOps pipeline variables / service connections for:

• Container registry credentials (ACR or Docker Hub)
• Kubernetes / Git credentials (if pipelines push manifest updates to the manifests repo)

Continuous Delivery with ArgoCD

ArgoCD continuously monitors a Git repo and a path (or a Helm chart) and reconciles a Kubernetes cluster to match the declared manifests. Typical flow with this repo:

1. CI builds and pushes a new image tag.
2. CI updates the Kubernetes manifest (image tag) using scripts/updatek8smanifest.sh and commits the change to the manifests branch.
3. ArgoCD detects the Git change and syncs the new manifests to the cluster, causing a rolling update of the deployment.

You can also use image automation tools (like ArgoCD Image Updater or Flux image automation) to avoid manual manifest commits.

Kubernetes

This project expects Kubernetes manifests compatible with the example-voting-app (Deployments, Services, ConfigMaps). The scripts/updatek8smanifest.sh is provided to help substitute new image tags into manifests before committing.

Common kubernetes commands used during debug or local testing:

# apply manifests
kubectl apply -f ./k8s-manifests/

# view deployments
kubectl get deployments -n <namespace>

# view pods
kubectl get pods -n <namespace>

# logs
kubectl logs -l app=vote -n <namespace>

Output

The below images show the output of the CI/CD pipeline