Trying to use Traefik does not seem to be easy with GUI containers as names are random, and no labels are exposed in the setting.
Trying to use Ngnix Proxy Manager does not seem to be easy either, as ZimaOS binds to all IP addresses on the NIC, so you would need a secondary adapter to make that work.
Is ZimaOS just not meant for reverse proxy setups? Is there an alternative setup I am just not aware of?
Yes, reverse proxy is possible on ZimaOS, but I agree it is not as clean as a normal Docker Compose setup where you fully control labels, service names, networks, and ports.
For Traefik, the main problem is exactly what you said. ZimaOS GUI-installed apps do not expose an easy way to add Docker labels, and the container names can be awkward/random, so Traefik loses one of its biggest advantages: automatic discovery.
Nginx Proxy Manager is usually the easier option on ZimaOS, but you need to plan around ports 80 and 443, because ZimaOS itself may already be using them. If NPM cannot bind to those ports, you either need to move the ZimaOS web UI to another port, use different external ports, or use another network/interface setup.
The cleaner method is usually:
- Run Nginx Proxy Manager as a normal Docker app.
- Put NPM and the target containers on the same Docker network where possible.
- Proxy to the container name and internal app port, not to the ZimaOS host IP.
- Only expose
80 and 443 through NPM.
- Avoid pointing NPM back to the ZimaOS IP unless you specifically want the ZimaOS UI.
So no, I would not say ZimaOS is “not meant” for reverse proxy setups. It can do it. But the current GUI app model makes advanced reverse proxy setups less straightforward than plain Docker Compose or Portainer.
In my opinion, Nginx Proxy Manager is the most realistic option for most ZimaOS users. Traefik can work, but it is better suited when the stack is deployed manually with Compose so you can define the labels and networks properly.
1 Like