CRITICALNginxNetworking

Nginx returns 502 Bad Gateway under load

nginxproxy502networkingperformance

Symptoms

Nginx error log shows 'upstream prematurely closed connection'
502 rate spikes in sync with request volume
Backend service logs look healthy

Root Cause

keepalive_requests on upstream exhausted and Nginx re-opens connections aggressively
Backend timeout is shorter than Nginx proxy_read_timeout
Backpressure from a slow dependency

Diagnosis

tail -f /var/log/nginx/error.log | grep upstream
ss -s on the Nginx host - look at TIME_WAIT count
Compare upstream_response_time vs request_time

Fix

Tune upstream keepalive:

upstream app {
  server app:8080;
  keepalive 64;
  keepalive_requests 1000;
  keepalive_timeout 60s;
}
server {
  location / {
    proxy_http_version 1.1;
    proxy_set_header Connection "";
    proxy_read_timeout 30s;
  }
}

Ensure backend keepalive timeout > Nginx keepalive_timeout

Prevention

Load test before launch and alert on 5xx > 1%
Emit upstream_response_time to Prometheus
Use circuit breakers in the application tier