AWS Compute

In production, AWS Elastic Compute Cloud (EC2) instances are treated as disposable compute nodes. Their primary storage is decoupled using **Elastic Block Store (EBS)** SSD volumes, allowing compute instances to fail or terminate without losing persistent database or log files.

Deploying in a single Availability Zone represents a critical single point of failure (SPOF). Production services must use **Auto Scaling Groups (ASGs)** to provision and balance EC2 instances dynamically across multiple AZs.

ASGs rely on **Launch Templates**—which define the AMI, instance size (e.g. m5.large), security groups, IAM instance profiles (to securely authorize S3/KMS access without hardcoded keys), and a custom **User Data bootstrap script** that executes on boot to configure the application environment.

Trace the operational flow of how AWS ASGs handle scaling events and self-healing: when an instance fails its HTTP load balancer health checks, the scaling controller intercepts the state, terminates the degraded node, and triggers a launch template to bootstrap a replacement instance in an active zone.

Below is the exact production-ready Shell Script attached to the Launch Template. It automates package installation, pulls credentials from S3 using IAM instance profiles, and configures Nginx to serve the app securely:

#!/bin/bash
set -e
# Install dependencies
apt update && apt install -y nginx php8.2-fpm php8.2-mysql git unzip
# Download environment settings securely from S3 (Requires Instance Profile)
aws s3 cp s3://company-config/prod/.env /var/www/app/.env
# Setup Nginx VHost
cat > /etc/nginx/sites-available/app << 'EOF'
server {
    listen 8080;
    root /var/www/app/public;
    index index.php;
    location / { try_files $uri $uri/ /index.php?$query_string; }
    location ~ \.php$ {
        fastcgi_pass unix:/var/run/php/php8.2-fpm.sock;
        include fastcgi_params;
    }
}
EOF
ln -sf /etc/nginx/sites-available/app /etc/nginx/sites-enabled/default
systemctl restart nginx