Wrong virtual host mapping after server reboot (domains showing incorrect site)

[cod3rshotout]

Description

Hi,
I’m using nginx-proxy-automation on a Linux server running Docker, and I’ve noticed an intermittent issue:
after a full server reboot, some domains occasionally serve the wrong website, meaning the vhost mapping becomes incorrect.

Example scenario:

• site-a.com unexpectedly shows the content of site-b.com
• Restarting the affected container or reloading nginx fixes the issue temporarily

This doesn’t happen after every reboot, only sometimes.

---

Expected behavior

Each domain should always route to the correct container after the host system reboots.

---

Actual behavior

After reboot:

• Wrong container is served for one or more VIRTUAL_HOST domains
• It looks like nginx-proxy loads an outdated or incomplete configuration
• Running nginx -s reload inside the proxy container fixes the issue
• Restarting the app container also fixes it

---

Possible causes (hypothesis)

• Race condition: proxy starts before all containers are alive
• Old config not fully refreshed on startup
• Docker network not ready when proxy boots
• Exited/ghost containers being detected by nginx-proxy
• Missing forced reload when containers attach to the proxy network

---

What I’ve tried

• Added restart: always to all containers
• Ensured all containers have unique VIRTUAL_HOST, LETSENCRYPT_HOST, VIRTUAL_PORT
• Added depends_on so apps start before proxy
• Running docker exec nginx-proxy nginx -s reload solves it
• Cleaned up stopped containers (docker container prune)
• Checked Docker networks (no duplicates or recreations)

---

Questions

1. Is this a known issue related to container startup order?
2. Is there a recommended way to force nginx-proxy to wait until all services are up?
3. Should an automatic nginx -s reload be triggered after boot?
4. Are there debug logs or flags that can help identify why config generation fails?