nginx 504 Gateway Timeout: Slow Upstream or Wrong Timeout

What you'll see

• 504 on long-running requests (uploads, reports, AI endpoints)
• Fast requests on the same site work fine

Root causes

Fix it

1. Find which requests time out and how long they take

   tail -f /var/log/nginx/access.log | grep ' 504 '

2. Raise timeouts honestly for the slow route only

   location /reports/ {
     proxy_read_timeout 300s;
     proxy_send_timeout 60s;
   }

3. If it's saturation, check upstream concurrency

   upstream app { server 127.0.0.1:8080 max_conns=50; }

4. For anything minutes-long, move to a job queue instead

   # return 202 + job id; let a worker do the work; poll or push the result