Latest deployment was missing internal entrypoint in traefik config

[theCalcaholic]

Describe the bug I recently updated my instance to commit 3e57c9f6e and noticed that all bridges were broken (I'm mainly using mautrix-whatsapp and mautrix-telegram).

During my analysis I finally found that traefik/config/traefik.yml was missing the following entrypoint (I found a warning in the logs that it was missing and added it manually following the docs for self-hosted traefik servers, which I'm not using):

matrix-internal-matrix-client-api:
  address: :8008
  forwardedHeaders:
    insecure: true

This fixed my issue, but I still don't know why the entrypoint wasn't generated by the playbook.

Let me know if I can provide any more information to help debugging.

My Setup

The most relevant part about my setup is probably the (reverse) proxy setup, so I'll describe it shortly:

I'm running an nginx reverse proxy fronting the traefik proxy from the playbook. The external endpoints were all working, the issue was solely with containers connecting via the traefik container directly (using port 8008). Certificate management is offloaded to the nginx proxy, as well as tls termination.

[enekonieto]

@theCalcaholic's workaround solved the issue in my homeserver, same setup as them (nginx reverse proxy fronting the traefik proxy from the playbook). Thanks!