Traefik loses all bridge network connections when modifying port mappings or toggling dashboard

[that0n3guy]

To Reproduce

1. Have multiple Docker Compose services deployed with domains (using isolated bridge networks)
2. Go to Web Server → Traefik settings
3. Either:
   • Enable or disable the Traefik dashboard, OR
   • Click "Save" on "Additional Port Mappings" (even without making changes)
4. All compose services become inaccessible (404 errors)

Additionally, when performing these actions in the UI, a "Failed to fetch" error appears in the lower right corner.

Workarounds

Option 1: Manually reconnect Traefik to all compose networks:

for net in $(docker network ls --filter "driver=bridge" --format "{{.Name}}" | grep "^saas-"); do
  docker network connect "$net" dokploy-traefik 2>/dev/null && echo "Connected: $net"
done

Option 2: Redeploy each affected compose service via the UI.

What I've already tried

1. Ran the breaking changes fix from v0.25.7 release notes:

docker pull traefik:v3.6.1
docker stop dokploy-traefik || true
docker rm dokploy-traefik || true
docker run -d \
  --name dokploy-traefik \
  --restart always \
  -v /etc/dokploy/traefik/traefik.yml:/etc/traefik/traefik.yml \
  -v /etc/dokploy/traefik/dynamic:/etc/dokploy/traefik/dynamic \
  -v /var/run/docker.sock:/var/run/docker.sock \
  -p 80:80/tcp \
  -p 443:443/tcp \
  -p 443:443/udp \
  traefik:v3.6.1
docker network connect dokploy-network dokploy-traefik || true

2. Ran update script: curl -sSL https://dokploy.com/install.sh | sh -s update

Issue persists after both attempts.

Related Issues

• dokploy-traefik container loses all docker networks when adjusting additional port mapping (removing duplicate traefik dashboard port mappings) #3022 (duplicate of Traefik "Enable dashboard": dokploy traefik container gone, all services domains down. [with solution] #2996)
• Traefik "Enable dashboard": dokploy traefik container gone, all services domains down. [with solution] #2996 (Traefik dashboard breaks networks)
• v0.25.7 update made all my apps to be unavailable to traefik #3053 (v0.25.7 update made apps unavailable)

Current vs. Expected behavior

Current: When Traefik container is recreated/restarted via the UI, it loses all bridge network connections to compose services. Only dokploy-network remains connected. Services return 404 until manually reconnected or redeployed.

Expected: Traefik should maintain (or automatically reconnect to) all compose bridge networks after restart/recreation.

Debugging

Verify missing networks:

for net in $(docker network ls --filter "driver=bridge" --format "{{.Name}}" | grep -E "^saas-|^[a-z]+-[a-z]+-[a-z0-9]{6}$"); do
  if ! docker network inspect "$net" --format '{{range .Containers}}{{.Name}} {{end}}' | grep -q "dokploy-traefik"; then
    echo "Missing traefik: $net"
  fi
done

Provide environment information

OS: ubuntu 24.04
dokploy version v0.26.2
VPS Hetzner
All these apps are docker compose deployments.
- **Traefik version:** v3.6.1

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

Traefik

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

Same server where Dokploy is installed

Additional context

No response

Will you send a PR to fix it?

No

[Siumauricio]

I tested the same conditions and it works fine check the video

Screen.Recording.2025-12-20.at.12.23.29.AM.mov

[Siumauricio]

I tested using isolated and not isolated, by any chance are you using docker stack?

[that0n3guy]

@Siumauricio I'm not using docker stack.

Your video shows you clicking the "reload" button on traefik... but not adding "additional port mappings" or enable/disabling traefik dashboard. Those are the elements that kills it... I don't think the reload button kills it for me.

Also, my apps are docker compose apps... I couldn't tell in your video how you had wordpress setup.

[eapronin]

I've also encountered this issue. If I use the domain, it gives me an error, but if I use the IP address, everything works fine

P.S. the video doesn't attach for some reason.

[Dappner]

I thought I had the same issue. It was actually Traefik behavior: it won't forward traffic until health checks pass.
In my case, the health check on my entrypoint (nginx) was failing. Dokploy didn’t surface it in the UI, so I only noticed after SSH’ing into the host. Once I fixed the health check, Traefik started routing correctly.

[nathaliedpuc]

For me I had added tailscale domains and toggled the dashboard. It knocked all the services that were related to tailscale domains off dokploy. It did not stop the compose but cleared out tailcale domain resolution networking

EDIT. Temporary workaround when this happens is to re add the running containers back to the dokploy network using
docker network connect dokploy-network <container-name>

[Siumauricio]

Okay, thank you for sharing the reproduction. It's a serious problem. We definitely need to reconnect all containers when recreating the container.

[tasiotas]

can this reconnect feature also run on startup?
my containers with Isolated Deployments arent reconnected to traefik after vm reboots.

im on 0.29.11