How do I fix 502 Bad Gateway with Nginx reverse proxy to a private backend?

[Brian.W]

I’ve been trying to set up a reverse proxy for my home lab server to access a few web apps securely, but I keep running into 502 Bad Gateway errors. I’m using Nginx on a VPS, and my backend is on a different private IP, but the connection just won’t stay stable.

[LoganL]

That sounds familiar. Sometimes the 502 shows up even when nginx can reach the host, because the backend closes the connection or returns a bad gateway while it's still starting.

[JerryR]

DHCP on the backend caused the private IP to drift; once I assigned a static private IP and pinned a route, the proxy stopped flaking.

[KennethXS]

I once pointed proxy_pass at a port the backend wasn't listening on; a quick curl showed it, and switching to the real port cured the 502.

[Brian43]

Question: are you sure the VPS can actually reach the private IP consistently? Could firewall rules be blocking the traffic in certain hours?

[Aria_W]

TLS termination mismatch can cause 502 too; if nginx is talking TLS to a plain HTTP backend, you get that kind of error.

[Avery84]

I tried longer timeouts and enable keepalive, but the 502 would still show up during spikes and it turned out the backend process was crashing under load.

[AubreyYJ]

I eventually isolated a small internal proxy in front of the backend just to sanity-check path; it stayed stable, so the issue looked like something in the path from the VPS to the LAN.