mirror of
https://github.com/fosrl/pangolin.git
synced 2026-05-26 16:21:17 -05:00
[GH-ISSUE #618] Cannot Proxy Docker Containers Running Alongside Pangolin in the VPS #1531
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: github-starred/pangolin#1531
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @epoch-philosophy on GitHub (Apr 29, 2025).
Original GitHub issue: https://github.com/fosrl/pangolin/issues/618
Hey all, this may be a pretty dumb one. But, I have a single VPS that houses Pangolin. It also contains some monitoring containers like Ntfy and Uptime-Kuma. I keep those on a VPS because my cat likes to turn off my local machines here and there and still want to get notifications in case she turned off the server housing Uptime-Kuma! (She's done this many times.)
The problem:
I cannot get Pangolin to properly proxy Uptime-Kuma (or any other local docker container on the VPS). I've tried via the "local" site connection, and even deployed a Newt connection on the same VPS. No avail.
I made sure to have the Uptime-Kuma container on the Pangolin network, used the public IP as the http source. I used it's docker IP, it's docker host name, and everything else I can think of.
I have a feeling I may need to route internal containers through Traefik as an external route/service in the dynamic_config.yml, but wanted to check first here, as there's no official documentation.
(Also know it's probably best practice to have a dedicated VPS for Pangolin.)
@miloschwartz commented on GitHub (Apr 30, 2025):
You can try to use
172.17.0.1to address the host of the machine and make sure Uptime-Kuma has a port bound to the host or put the container in the same compose as Pangolin and address it by the container name .@epoch-philosophy commented on GitHub (Apr 30, 2025):
I tried that. Did not work. Also, should that be on a Newt connection, or local? I think the only thing about putting Uptime-Kuma in the same compose file is that they would share the same docker network. Which you can still do (what I specifically did) in a separate compose file.
Has anyone ran into this issue or had to use some pretty unorthodox things to proxy it? My only thought would be to run the container normally, and expose the 3001(Uptime-Kuma's) port on the VPS, then run it as an external service in the dynamic_config.yml to be proxied that way. Similar to how Traefik would proxy something outside the host. Problem with this is, is that's a public facing IP and it exposes Uptime-Kuma to the internet via unencrypted HTTP. Which isn't great.
Is there a more elegant and safe way to proxy local docker containers on the VPS? (If so, that would be a great thing to add to documentation in the future.)
@v1rusnl commented on GitHub (Apr 30, 2025):
Local connection, setup e.g. kuma connected to the pangolin network, address it with container name and port (you need to address the internal container port here, not the port to the outside) in the Pangolin UI is the correct way.
If that's not working there is something else going on on you VPS.
@epoch-philosophy commented on GitHub (Apr 30, 2025):
That did it! Weirdly enough I did EXACTLY that at one point and I think my browser was caching the HTTP error/bad gateway errors I was receiving and I didn't clear it. Thank you for the solution.
There's a section in the documentation around the danger of Docker opening ports automatically on the VPS when running containers alongside Pangolin, perhaps this would be a good place to document running Docker containers in the same VPS! (If you are open to edits, I'd be happy to help.)