Troubleshooting an unreachable Linux cloud server: five layers to check

A server responding on localhost may still be inaccessible from the internet, so each layer should be tested separately rather than resizing the instance. Administrators are advised to check which ports are listening and whether a backend bound only to 127.0.0.1 is intentionally kept private behind a reverse proxy like Nginx. The operating-system firewall and cloud instance inbound rules should be reviewed, opening only required public ports and restricting SSH to trusted addresses. For proxy issues, a 502 error points to checking the upstream address, port, application status and error logs, while wrong-site responses may stem from server_name or stale DNS records. The guide also recommends verifying certificates against the hostname, testing from external networks, and keeping a redacted record of test results.