dokploy-traefik container loses all docker networks when adjusting additional port mapping (removing duplicate traefik dashboard port mappings)

[josias-r]

To Reproduce

1. Enable treafik dashboard (not sure this is necessary, but the ports added there started my confusion)
2. Create applications and assign domains. Use isolated networks. Deploy it.
3. Application access works fine via traefik reverse proxy
   • Docker inspect networks shows that the dokploy-traefik container has the application network assigned as well, as expected
4. Go to web server additional port mapping settings
5. remove one of the 8080 mappings
6. Dokploy will restart dokploy-treafik container.
7. Treafik cannot access application anymore
   • docker inspect shows, dokploy-treafik container only has one network attached now. all other networks are gone. this is the reason why traefik will gateway timeout on ALL applications now.

Current vs. Expected behavior

Removing the duplicate port mapping (8080 auto created by enabling treafik dashboard) does not break all docker networks that were automatically set-up.

Current: duplicate port mapping, removing one restarts treafik. breaks networks, and duplicate port mapping remains

Expected: no duplicate port mapping, removing it WONT break networks.

Provide environment information

- Ubuntu 22.04.5 LTS
- x86_64
- Dokploy 0.25.6

Which area(s) are affected? (Select all that apply)

Traefik, Docker, Application

Are you deploying the applications where Dokploy is installed or on a remote server?

Same server where Dokploy is installed

Additional context

I stumpled upon this by accident and investigated it. All I did was browse the dokploy dashboard and check the settings.

I was confused that there is a duplicate 8080 tcp port mapping and was not sure where it came from. I just removed one of them and pressed save. Only later I realized that these were auto added from enabling the treafik dashboard. (i dont know where there is a duplicate mapping still tho)

I would expect to update the additional port mappings would NOT remove/break all networks setup by dokploy and treafik

Will you send a PR to fix it?

No

[dsincl12]

I just encountered this as well, added a port mapping and removed it now everything is completely broken.

[josias-r]

I might add, that re-deploying each service will re-add the networks to the traefik container, aka fix the access again

[dsincl12]

ok much appreciated information, still can't HTTPS working again, not for services or the control panel either :(

[dsincl12]

ok I got it sorted, my issue was that I had a redpanda-console running on port 8080 in docker already (separate from Dokploy), when I enabled the traefik dashboard that killed traefik completely with:

2025-11-15T11:56:20.286578139Z [Error: Command failed: docker restart dokploy-traefik
2025-11-15T11:56:20.286589791Z Error response from daemon: Cannot restart container dokploy-traefik: container e7e425a1f7aafd4d0aac955eeaffb72dced215c1075b9a8d12f0ef828b1b0099: endpoint join on GW Network failed: driver failed programming external connectivity on endpoint gateway_b1a1eca88215 (5f98f4dfb2b5f08cfa114f16290715a826e55ce9bf3187a7cb872f0baa5148c9): Bind for 0.0.0.0:8080 failed: port is already allocated
2025-11-15T11:56:20.286593238Z ] {
2025-11-15T11:56:20.286595652Z code: 1,
2025-11-15T11:56:20.286597876Z killed: false,
2025-11-15T11:56:20.286600040Z signal: null,
2025-11-15T11:56:20.286602245Z cmd: 'docker restart dokploy-traefik',
2025-11-15T11:56:20.286604449Z stdout: '',
2025-11-15T11:56:20.286606663Z stderr: 'Error response from daemon: Cannot restart container dokploy-traefik: container e7e425a1f7aafd4d0aac955eeaffb72dced215c1075b9a8d12f0ef828b1b0099: endpoint join on GW Network failed: driver failed programming external connectivity on endpoint gateway_b1a1eca88215 (5f98f4dfb2b5f08cfa114f16290715a826e55ce9bf3187a7cb872f0baa5148c9): Bind for 0.0.0.0:8080 failed: port is already allocated\n'
2025-11-15T11:56:20.286609358Z }
2025-11-15T11:59:11.981162070Z  ELIFECYCLE  Command failed.

So I had to stop Redpanda Console in Docker, restart Traefik container to get everything up and running again, disable Traefik dashboard, and finally start everything up again normally.

[kheiner]

Does #2997 address the issue here?

[dsincl12]

for my case yes, that would've saved me

[superadmin7]

• Dokploy 0.25.6

I'm facing an issue when running Traefik with Docker Engine 29.

Traefik keeps repeating this error:

"Error response from daemon: client version 1.24 is too old. Minimum supported API version is 1.44"

However, my Docker Engine exposes API version 1.52 (well above the minimum).

It seems that Traefik (or the Swarm provider) is using an outdated internal Docker client (API v1.24), which is incompatible with Docker 29.

As a result, Traefik cannot connect to the Docker daemon and cannot load the providers.

Screenshot attached.

Is there a fix or a Traefik version that uses a newer Docker API client (>= 1.44)?

[Bima42]

@superadmin7 This PR will create a new release that will fix this issue
You should update traefik by yourself, you can look here if you don't know how to do it