The Root Cause

Nginx 502 Bad Gateway on AWS EC2 typically indicates Nginx is unable to communicate with its designated upstream application server (e.g., PHP-FPM, Gunicorn, Node.js process). This often stems from the application process crashing due to resource exhaustion (like out-of-memory errors) specific to the EC2 instance’s size, or the application failing to start or bind to the expected port/socket.

Quick Fix (CLI)

# 1. Identify and check the status of your upstream application service.
#    Replace 'your-application-service' with the actual service name (e.g., php-fpm, gunicorn, uwsgi, node-app).
#    Example for PHP-FPM (versions vary, e.g., php7.4-fpm):
#    sudo systemctl status php-fpm
#    Example for a Gunicorn service managed by systemd:
#    sudo systemctl status gunicorn

# 2. Review logs for the application service for specific errors (e.g., OOM, unhandled exceptions).
#    sudo journalctl -xeu your-application-service -n 50

# 3. Attempt to restart the upstream application service.
#    sudo systemctl restart your-application-service
#    For PM2-managed Node.js apps: pm2 restart all

# 4. Restart Nginx to ensure it re-establishes connections to the upstream service.
sudo systemctl restart nginx

Configuration Check

Review your Nginx server block configuration, typically found in /etc/nginx/conf.d/*.conf or /etc/nginx/sites-available/*. Focus on the proxy_pass or fastcgi_pass directives within your location blocks to ensure they point to the correct address/port/socket where your application server is listening.

# Example for PHP-FPM (fastcgi_pass):
# Verify this path matches your PHP-FPM pool configuration (e.g., /etc/php/7.4/fpm/pool.d/www.conf 'listen' directive).
location ~ \.php$ {
    include fastcgi_params;
    fastcgi_pass unix:/run/php/php7.4-fpm.sock; # Ensure this socket path is correct
    # OR (if using TCP)
    # fastcgi_pass 127.0.0.1:9000; # Ensure this IP and port are correct
    fastcgi_index index.php;
    fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
}

# Example for general HTTP upstream (proxy_pass, e.g., Gunicorn, Node.js):
# Verify this IP and port match your application's 'bind' address.
location / {
    proxy_pass http://127.0.0.1:8000; # Ensure this IP and port are correct
    proxy_set_header Host $host;
    proxy_set_header X-Real-IP $remote_addr;
    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
}

Additionally, confirm your upstream application’s configuration (e.g., /etc/php/7.4/fpm/pool.d/www.conf’s listen directive for PHP-FPM, or your Gunicorn/Node.js app’s bind address) matches the Nginx configuration.

Verification

# 1. Test Nginx configuration for syntax errors.
sudo nginx -t

# 2. If the test is successful, reload Nginx to apply any configuration changes.
sudo systemctl reload nginx

# 3. Verify the website is now accessible by making an HTTP request.
curl -I http://localhost
# OR (if testing from outside the EC2 instance, or for a specific domain)
curl -I http://your-ec2-public-ip-or-domain.com