Fixing “Web server is down (521)” When Using Nginx + Cloudflare

Recently I hit a 521 error from Cloudflare even though my server was running. Here’s the step-by-step process that solved it.

1️⃣ Check Nginx Status

sudo systemctl status nginx

I found:

[emerg] host not found in upstream "ip-10-0-0-25.internal"

This means Nginx could not resolve the upstream host in its config.

2️⃣ Locate the Problem Config

sudo nano /etc/nginx/sites-enabled/example-site.conf

Inside I had:

proxy_pass http://ip-10-0-0-25.internal:3000;

The hostname was not resolvable.

3️⃣ Replace With a Private IP or Stable DNS

Find the backend’s private IP (example):

hostname -I
# → 10.0.1.42

Update the config:

proxy_pass http://10.0.1.42:3000;

Save and exit (Ctrl+X, Y, Enter).

4️⃣ Test and Reload

sudo nginx -t
sudo systemctl restart nginx

🔑 Tips

• Use Elastic IP or a private DNS name if the backend instance’s private IP might change.
• Keep your firewall/security group rules tight: allow port 3000 only from the Nginx proxy server.
• Always run nginx -t before reloading to catch syntax errors.

💡 Summary A 521 error often means Cloudflare can’t reach your origin. In my case the root cause was an invalid hostname in the Nginx proxy_pass. Switching to a valid private IP (or a resolvable internal DNS name) fixed it immediately.