External App Proxy Setup Causing Nginx Auto-Restart Help with Reverse Proxy Architecture?

[markpousee]

Hi Cloudron community,

I’ve hit a bit of a snag while working through a custom setup involving an external reverse proxy and Cloudron app deployment.

Here’s what I’ve done so far:

• Configured a second VPS running an external reverse proxy (via Nginx Proxy Manager) to forward certain subdomains to my Cloudron instance on a primary VPS.
• Followed a pattern similar to the "Deploying Anubis (AI Crawler Filtering)" setup creating Docker Compose stacks on the proxy server, including a Redis container for session and metrics support.
• Used REDIRECT_DOMAINS in Anubis (the proxy component) to map Cloudron subdomains (e.g., app1.example.cloud, app2.example.cloud) properly.

Everything appears correct, but the issue is: Nginx keeps restarting intermittently on the proxy server causing 502 / gateway errors when users hit certain Cloudron-hosted apps.

Here’s what I’ve tried:

• Double-checked the nginx.conf for syntax errors nginx -t shows no problems.
• Validated SSL certificate settings for *.example.cloud, with Let’s Encrypt configured and working.
• Confirmed Cloudron apps are healthy and listening correctly when accessed directly (not via the proxy).
• Ensured both VPS firewalls (UFW on VPS1 and VPS2) allow traffic on TCP 443 and 80, plus internal management ports.

A few questions I’m hoping the community can help with:

• Has anyone seen Nginx auto-restarting in this route-proxy setup, and what diagnostics helped (e.g., journalctl, Docker resource limits)?
• When routing Cloudron subdomains through a proxy, are there specific headers, health checks, or Nginx tuning parameters that are essential?
• Is there a better architecture approach like using Cloudron’s built-in reverse proxy instead or a more crash-resilient pattern for handling incoming traffic?

I want to nail down a stable and secure proxy layer without disrupting Cloudron’s internal routing.

Thanks in advance for any insights or experiences you can share!

Mark Pousee

[joseph]

@markpousee have you already set the proxy server as a Trusted IP on cloudron - https://docs.cloudron.io/networking/#trusted-ips ?

Though, I cannot think of why nginx is automatically restarting . TBH, I have never seen nginx crash. This makes me wonder if something in your setup (compose?) is monitoring nginx responses (some healtcheck?) and restarting nginx automatically ?

If you have a test Cloudflare, you can quickly put Cloudflare in front of Cloudron and check if that works. After all, your proxy is same as Cloudflare functionally.