GHOSTVPS / Operations

Find why a domain, HTTPS or website returns an error

Trace DNS, TLS, reverse proxy and application layers, including a real isolated 502-to-200 recovery.

Updated · GhostVPS

Before you start

Control of a domain and a web service you administer. Replace example.com, APP_PORT and YOUR_APP_SERVICE with your own values. The 502/200 images come from an isolated NGINX listener on localhost, not the live GhostVPS website.

Identify the failing layer before changing DNS

Record the exact URL, error code, time and whether the problem affects one client or everyone. Query A/AAAA and nameserver records; if you use a proxy such as Cloudflare, the public A record may show proxy addresses rather than your origin IP. Compare public DNS with the intended provider configuration. Do not repeatedly change records while a previous change is still propagating.

Your computer · terminal

dig example.com A +short
dig example.com NS +short
Real public DNS and HTTPS response checks recorded during the lab.
Real public DNS and HTTPS response checks recorded during the lab.

Check the HTTP and HTTPS response

Inspect response headers and status for both the page and a known-good static asset. A TLS certificate error, DNS failure, HTTP 404 and HTTP 502 point to different layers. Our captured DNS/HTTPS check received a successful HTTP 200 from the live site; it does not imply your origin will be healthy.

Your computer · terminal

curl -I https://example.com/

Read the reverse proxy error before restarting

Validate the NGINX configuration and inspect its error log around the failed request. In the isolated lab, NGINX forwarded to a nonexistent local application and returned a genuine 502. A 502 can mean the upstream refused a connection, timed out or returned an invalid response; the error log gives the specific reason.

Your VPS · Linux terminal

sudo nginx -t
sudo tail -n 50 /var/log/nginx/error.log
The isolated NGINX reverse proxy returned HTTP 502 for an unavailable upstream.
The isolated NGINX reverse proxy returned HTTP 502 for an unavailable upstream.

Test the application behind the proxy

Check that the upstream process is running and listening at the host and port configured in the proxy. Request it directly from the server using its local address. In the lab, correcting the isolated upstream path produced HTTP 200 without touching public DNS. On a real site, fix the application or proxy target indicated by logs, then test configuration before reload.

Your VPS · Linux terminal

curl -I http://127.0.0.1:APP_PORT/
systemctl status YOUR_APP_SERVICE --no-pager
After fixing only the isolated lab configuration, the local service returned HTTP 200.
After fixing only the isolated lab configuration, the local service returned HTTP 200.

Verify recovery from inside and outside

Repeat the local upstream request, public HTTPS request and a browser check from another network. Confirm redirects, certificate validity and content, not just status code. If the fix fails, restore the previous known-good configuration and use the recorded error to continue diagnosis.

Official references