☁️ Uptime Kuma — Cloud & Kubernetes Deployment

Forked from louislam/uptime-kuma | Production-ready infrastructure built on top.

This fork focuses entirely on the infrastructure layer: deploying Uptime Kuma on AWS using multiple orchestration strategies — from serverless Fargate to a fully managed EKS cluster with Helm.

---

🚀 Deployment Strategies

1. AWS Fargate + EFS + SNS (Serverless)

Serverless deployment with persistent storage, remote state, and automated alerting.

Key features:

• ECS Fargate task (no EC2 to manage)
• EFS mount at /app/data for persistent SQLite data
• SNS + CloudWatch for CPU monitoring and email alerts
• VPC with 2 public subnets across Availability Zones
• Terraform remote state in S3 (versioned, AES-256 encrypted, public access blocked)
• GitHub Actions CI/CD: fmt, validate, init -backend=false on every push

🏗 Architecture

📸 Live Dashboard

(Active monitoring dashboard showing real-time service status)

---

2. Kubernetes Manifests (Bare)

Full K8s deployment from scratch using kubectl.

Resource	Purpose
Deployment	Uptime Kuma pod spec
Service	ClusterIP internal routing
PersistentVolumeClaim	Persistent data storage
ConfigMap	Environment configuration
Secret	Sensitive config (base64)
Ingress	External HTTP access

k8s/manifests/

3. Helm Chart (Custom)

Custom-authored Helm chart for parameterized, multi-environment deployments.

• Configurable via values.yaml
• Templated Deployment, Service, Ingress, PVC

helm install uptime-kuma ./k8s/helm -n monitoring

4. EKS Cluster on AWS (Terraform)

Live EKS cluster provisioned in eu-central-1 via Terraform. Deployed and validated end-to-end.

cd k8s/terraform && terraform init && terraform apply
aws eks update-kubeconfig --region eu-central-1 --name uptime-kuma-cluster
helm install uptime-kuma ./k8s/helm -n monitoring

---

📁 Repo Structure

uptime-kuma/
├── .github/workflows/
│   └── terraform-ci.yml      # CI/CD: fmt + validate on every push
├── img/
│   ├── kuma-arc.drawio.png   # Architecture diagram
│   └── kumademo.png          # Dashboard screenshot
├── terraform/
│   ├── bootstrap/            # S3 bucket for remote state (run once)
│   ├── backend.tf
│   └── ...
├── k8s/
│   ├── manifests/            # Raw Kubernetes YAML files
│   ├── helm/                 # Custom Helm chart
│   └── terraform/            # EKS cluster (eu-central-1)
└── [original uptime-kuma source]

---

⚙️ CI/CD Pipeline

GitHub Actions runs automatically on every push to main and on pull requests.

Step	What it does
terraform fmt -check	Fails if code is not properly formatted
terraform init -backend=false	Initializes without connecting to S3
terraform validate	Checks for syntax and configuration errors

---

🛠 Quick Deploy

Fargate (AWS)

cd terraform/bootstrap && terraform init && terraform apply  # first time only
cd .. && terraform init && terraform apply
# Access: http://:3001

Kubernetes (local)

minikube start
kubectl create namespace monitoring
helm install uptime-kuma ./k8s/helm -n monitoring
kubectl port-forward service/uptime-kuma 3001:3001 -n monitoring

Kubernetes (EKS)

cd k8s/terraform && terraform init && terraform apply
aws eks update-kubeconfig --region eu-central-1 --name uptime-kuma-cluster
kubectl create namespace monitoring
helm install uptime-kuma ./k8s/helm -n monitoring

---

📝 Lessons Learned

• EFS Connectivity: Resolved ResourceInitializationError by opening port 2049 in the Security Group for EFS communication.
• Network Routing: Configured Internet Gateways and Route Tables within the VPC module for public access.
• Remote State Bootstrap: S3 backend cannot provision itself — a separate bootstrap/ module is required before terraform init.
• CI/CD without credentials: terraform init -backend=false allows fmt/validate checks in GitHub Actions without AWS credentials.

---

🧠 Tech Stack

AWS ECS Fargate · EKS · EFS · SNS · CloudWatch · VPC · IAM · S3 Terraform · Kubernetes · Helm · Docker · GitHub Actions

---

→ LinkedIn · Portfolio