Issues configuring Caddy

[fscheps]

Hi all, I am having issues with my caddy file configuration and I don't get it how to fix it, already tried with help from Gemini and ChatGPT4o but no luck.

The error I am seeing on the logs is:
{"level":"info","ts":1718918029.8250892,"msg":"using provided configuration","config_file":"/etc/caddy/Caddyfile","config_adapter":"caddyfile"}
Error: adapting config using caddyfile: syntax error: unexpected token '404', expecting '}', at /etc/caddy/Caddyfile:15 import chain: ['']
{"level":"info","ts":1718918036.809421,"msg":"using provided configuration","config_file":"/etc/caddy/Caddyfile","config_adapter":"caddyfile"}
Error: adapting config using caddyfile: syntax error: unexpected token '404', expecting '}', at /etc/caddy/Caddyfile:15 import chain: ['']

When I do a "sudo docker ps" I shows caddy restarting.

CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
f8ce9baec01a caddy:2.7-alpine "caddy run --config …" About an hour ago Restarting (1) 29 seconds ago root-caddy-1
92f4e11364e4 ghcr.io/defguard/defguard-proxy:latest "./defguard-proxy" 2 hours ago Up About an hour root-proxy-1
895e3dcccf52 ghcr.io/defguard/gateway:latest "/usr/local/bin/defg…" 2 hours ago Up About an hour root-gateway-1
abd1046d942f ghcr.io/defguard/defguard:latest "./defguard" 2 hours ago Up About an hour 0.0.0.0:50055->50055/tcp, :::50055->50055/tcp root-core-1
e64615e5021e postgres:15-alpine "docker-entrypoint.s…" 2 hours ago Up About an hour 5432/tcp root-db-1

This is how my Caddy File looks like:

`# Main domain with secure protocols
subdomain.domain.com {
tls admin@domain.com {
protocols tls1.3
}
# Reverse proxy configuration with additional headers
route / {
reverse_proxy core:8000 {
header_up X-Real-IP {remote_addr}
header_up X-Forwarded-For {remote_host}
header_up Host {http.request.host}
}
}
}

Management domain configuration

subdomain.domain.com {
tls admin@domain.com {
protocols tls1.3
}
route / {
reverse_proxy proxy:8080 {
header_up X-Real-IP {remote_addr}
header_up X-Forwarded-For {remote_host}
header_up Host {http.request.host}
}
}
}

HTTP to HTTPS redirection

http:// {
redir https://{host}{uri} permanent
}

Global error handling for 404 errors

handle_errors {
rewrite * /404.html
}

Global fallback for unspecified routes

:80, :443 {
respond "Not found" 404
} `

Thanks a lot for your help. I am sorry if I did something silly, I am not a Linux person and I am getting around by asking to ChatGPT but this one seems to be a tough cookie.

[fscheps]

The following post shares the paths I had to take to fix this issue. Thanks Teon for helping me on the Matrix support channel. Hope this helps anyone facing a similar issue:

Caddy Docker restarting in loop without letting me see any web interface of Defguard

While trying to deploy Defguard on a Debian 12 VM running in Proxmox, I was having the Caddy docker constantly restarting. When checking the logs I was seeing this:

Error: adapting config using caddyfile: syntax error: unexpected token '404', expecting '}', at /etc/caddy/Caddyfile:15 import chain: ['']
After reinstalling everything again with the one line script with the correct docker version installed as recommended by Teon. I was having another type of issue, which I asked ChatGPT to Describe as follows:

Issue 2: Caddy Configuration Causing Service to Restart

Problem: After setting up Defguard on a Debian 12 VM, the Caddy service responsible for reverse proxying was repeatedly restarting. Accessing the web interface resulted in seeing the internal routers home page instead of the intended Defguard interface.

Symptoms:
The Caddy container was restarting continuously.
Accessing the configured domains showed the internal routers home page.
Error logs indicated issues with multiplexing HTTP and HTTPS on the same port.
Error Message:
Error: adapting config using caddyfile: server listening on [:443] is configured for HTTP and cannot natively multiplex HTTP and HTTPS: https://example.com {
reverse_proxy 172.18.0.2:8000
}

https://admin.example.com {
reverse_proxy 172.18.0.3:8080
}

:80 {
respond 404
}
:443 {
respond 404
}

Resolution:
Correct the Caddyfile: Adjusted the Caddyfile to properly handle HTTP and HTTPS requests separately. The corrected Caddyfile included on-demand TLS configuration and redirection from HTTP to HTTPS.
Updated Caddyfile:
plaintext
Copy code
https://example.com {
reverse_proxy 172.18.0.2:8000
tls {
on_demand
}
}

https://admin.example.com {
reverse_proxy 172.18.0.3:8080
tls {
on_demand
}
}

:80 {
redir https://{host}{uri}
}
Enable Hairpin NAT: Enabled Hairpin NAT (NAT loopback) on the internal router to allow internal devices to access the external IP address.

Port Forwarding Configuration: Verified the port forwarding rules on the internal router to ensure ports 80 and 443 were forwarded to the internal IP address of the VM running Defguard.
Restart Caddy Container: Restarted the Caddy container to apply the changes.
docker restart root-caddy-1

Summary
The issue was due to improper handling of HTTP and HTTPS requests in the Caddy configuration file. By separating the handling of HTTP and HTTPS and enabling on-demand TLS, along with ensuring correct port forwarding and NAT loopback settings, the Defguard web interface was successfully accessed.
This solution should help others facing similar issues with Caddy configurations and network settings in a Defguard setup.

I have created a non official GPT to help specifically on DefGuard topics this is how I got to the solution, feel free to use it: https://chatgpt.com/g/g-uAJiWwPZD-defguard-assistant-not-official