Cloud Capstone | End-to-End Azure Deployment

You will deliver a production-shaped cloud application on Azure: infrastructure as code, a Kubernetes platform, a CI/CD pipeline, monitoring, and security controls. By the end the entire environment is reproducible from a Git repository.

Learning Objectives

Provision the platform with Terraform.
Deploy the app to AKS via CI/CD.
Add monitoring, alerts, and security guardrails.
Time: ~25 hours · Difficulty: Advanced · Prereqs: the Azure, AKS, IaC, and CI/CD labs.

Architecture Overview

graph LR Git[Git repo] -->|CI/CD| Pipe[Pipeline] TF[Terraform] --> Infra[VNet + AKS + DB] Pipe -->|deploy| AKS[AKS workload] AKS --> Mon[Azure Monitor] CSPM[CSPM scan] --> Infra

Environment Setup

Azure subscription, Terraform, kubectl, and a CI/CD runner.
A containerized app with tests.
Remote Terraform state in Azure Storage.

Step-by-Step Execution

Provision the platform with Terraform

terraform init && terraform apply -auto-approve

Creates VNet, AKS, and the database tier.

Deploy the app via CI/CD

git push origin main # pipeline builds, scans, and deploys to AKS

Verify monitoring and posture

$ kubectl get deploy && az monitor metrics alert list -o table

app 3/3 Available cpu-high Enabled

Progress So Far

graph LR A[01 IaC] -->|done| B[02 CI/CD deploy] B -->|done| C[03 Monitor + posture] style A fill:#1a4a1a,stroke:#00ff00,color:#fff style B fill:#1a4a1a,stroke:#00ff00,color:#fff style C fill:#1a4a1a,stroke:#00ff00,color:#fff

Testing & Validation

curl -s -o /dev/null -w "%{http_code}\n" https://app.lab.tyfsadik.org && terraform plan -detailed-exitcode

The app should return 200 and Terraform plan should show no drift. If both pass, the whole environment is reproducible and live.

Troubleshooting

Pipeline cannot reach AKS: grant the CI service principal the AKS cluster-user role.
Drift after deploy: ensure app changes go through CI, not manual kubectl edits.
Cost creep: destroy the environment when not in use (terraform destroy).

Extension Ideas

Add blue/green or canary deployments.
Integrate SOC monitoring for the workload.
Add a DR region and test failover.

Key Results

Delivered a full cloud app reproducible from Git.
Automated build, scan, and deploy to AKS.
Added monitoring, alerts, and posture scanning.
Confirmed zero drift between code and cloud.

Cloud Capstone: End-to-End Azure Application Deployment