Custom domain pointing to Coolify FastAPI server returns "no available server"

[Dijkie85]

Context

Hi everyone! I'm facing an issue while hosting a FastAPI application using Coolify on a VPS. Here’s the setup:

1. The Coolify dashboard is successfully running at https://coolify.mydomain.com.ar.

   • I have an A record in Cloudflare (proxied) pointing to my VPS IP, and it works perfectly.

2. My FastAPI service is healthy in Coolify, and I’ve set https://plus-api.mydomain.com.ar as the custom domain for the service.

   • I added a corresponding A record (proxied) in Cloudflare pointing to the same VPS IP.

Issue

When I try to access https://plus-api.mydomain.com.ar, I receive the error:
"No available server." (though it correctly connects via HTTPS)

---

Steps Taken So Far

Here’s what I’ve tried:

1. Verified Cloudflare’s SSL/TLS encryption is set to Full.
2. Checked that the DNS A record for plus-api.mydomain.com.ar correctly points to my VPS IP address.
3. Confirmed that the FastAPI service is healthy in the Coolify dashboard. Locally it's listening to port 8001
4. Ensured that ports 80 and 443 are open on the VPS (tested with ufw status).
5. Double-checked the custom domain configuration in Coolify and ensured it includes https:// for SSL handling.

---

Additional Details

• Coolify Version: Latest. Fresh install.
• Proxy: Traefik
• Cloudflare Settings: Proxy enabled (orange cloud), A record set to VPS IP.
• Error Message: "No available server."

---

Questions

1. Is there something specific I might be missing in the Coolify custom domain configuration?
2. Could Cloudflare’s proxied setup cause issues with Coolify’s proxy?
3. Are there logs in Coolify or somewhere else I should be checking to debug further?

Any advice or suggestions would be greatly appreciated. Thanks in advance for your help!

[Potapy4]

Have you tried explicitly specifying the FastAPI container port in your service?

Also, the logs can be found under Servers -> YOUR_SERVER -> Proxy tab -> Logs

[Dijkie85]

Thanks. I've tried that but it didn't work. What did work --which might be a bug-- was clicking off Readonly labels in the Network configuration. Somehow the actual Trafeik config was not updating until I disabled read only, and then it just worked.

[Dijkie85]

Yes, it seems to be a bug; I checked with another repo and the same happened: Trafeik config is not updating after changing the domain. The correct workaround is unticking the Escape special characters in labels?, not the Readonly. I'll submit the issue.

[minajevs]

Woop, thanks @Dijkie85! Stumbled on the same issue.