Express.js 502 Bad Gateway behind nginx — complete fix

TL;DR

A 502 Bad Gateway from nginx in front of Express.js means nginx received an invalid response from the upstream Express process — or no response at all. The Express process has either crashed, is too slow to respond, or is not listening on the expected address.

Common causes

• Express process crashed and has not restarted (check PM2 or systemd)
• Express is listening on a different port than nginx expects
• The proxy_pass target uses http://localhost:3000 but the app binds to a Unix socket (or vice versa)
• Upstream timeouts — Express takes longer than nginx's proxy_read_timeout
• The app ran out of memory and was killed by the OOM killer

Diagnosis workflow

Check if the Express process is alive:

pm2 status
# or
systemctl status my-express-app

Verify it is listening on the expected port:

ss -tlnp | grep :3000

Review nginx error logs for the specific upstream error:

tail -100 /var/log/nginx/error.log

Look for upstream prematurely closed connection, connect() failed, or upstream timed out.

Fix the nginx configuration

Ensure your upstream block matches where Express actually listens:

upstream express_app {
    server 127.0.0.1:3000;
    keepalive 64;
}

server {
    listen 80;
    server_name example.com;

    location / {
        proxy_pass http://express_app;
        proxy_http_version 1.1;
        proxy_set_header Connection "";
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_read_timeout 60s;
        proxy_connect_timeout 10s;
    }
}

Timeout alignment

Express, PM2, and nginx must agree on timeouts. If your Express route takes 30 seconds to respond, set proxy_read_timeout to at least 35 seconds. Ideally, move slow work to a background job queue so HTTP responses stay fast.

Where Reflex helps

Reflex correlates nginx 502 errors with upstream process health. When it detects a 502 spike, it checks whether the Express process is running, restarts it if needed, verifies the 502 rate drops, and alerts your team with a full incident timeline. See How it works.