middleware specified at entrypoint

[Xanderlicious]

I don't know if this has been mentioned already and I apologise if it has ( i've had a look but couldn't see anything that matches what I've noticed)

I have some headers applied at the entrypoint (defined in a dynamic config.yml file)

when viewing routes in traefik-manager, it only shows the middleware defined here for containers that are using traefik labels
for containers that use a dynamic file for their routes and service, they still have the headers applied but traefik-manager doesn't reflect this

[chr0nzz]

Thanks for reporting this! This is a known display gap - the middleware you've applied at the entrypoint level is actually working correctly in Traefik, it just doesn't show up in TM for routes that use config files. Routes using Docker labels show it because Traefik includes that info differently for those.

It's on my list to fix in a future release. Nothing broken on your end - your middleware is being applied, TM just isn't showing it.

[Xanderlicious]

`xander on titan at …/traefik-manager via  main

docker ps | grep traefik-manager
9251623ec13c ghcr.io/chr0nzz/traefik-manager:beta "gunicorn --bind 0.0…" 2 minutes ago Up 2 minutes (healthy) 0.0.0.0:5000->5000/tcp, [::]:5000->5000/tcp traefik-manager
🌀 xander on titan at …/traefik-manager via  main`

Just updated using the beta tag but it appears to be the same

[chr0nzz]

Thanks for testing the beta! For the entrypoint middleware display to work, TM needs access to your static config (traefik.yml). Make sure STATIC_CONFIG_PATH is set and the file is mounted into the container.

In your compose:

traefik-manager:
  environment:
    - STATIC_CONFIG_PATH=/etc/traefik/traefik.yml
  volumes:
    - /path/to/traefik.yml:/etc/traefik/traefik.yml:ro

Or you can set the path in Settings - Connection - File Paths without a restart. Once TM can read your static config, it will pick up any entryPoints[name].http.middlewares entries and show them as grey ep chips on the matching route cards.

[chr0nzz]

the beta image has added support for entrypoint middleware and if you want to test it for me use :beta tag and update

[chr0nzz]

Glad the middleware display is working!

For the domain issue - that's not TM, that'll be a Traefik routing/config issue. Check your Traefik logs after the container restart.
and show me

For the entrypoints - can you clarify where you're trying to specify them? In the Add/Edit Route form, entrypoints are selectable chips - just click each one you want, no commas needed. If you're asking about the ep middleware chips on route cards, those show automatically for any entrypoint that has middlewares defined in your traefik.yml - no extra config needed on TM's side.

[Xanderlicious]

I'll get back to you on this

I've ran out of time to deal with this further

but just to advise, I haven't change anything other than the compose file for traefik-manager

[chr0nzz]

try reverting back to :latest and see if it loads

[chr0nzz]

dont worry take your time and get back when ever u can thanks

[Xanderlicious]

So to deal with the domain issue,
i've reverted back to :latest and domain works :beta does not

with regards to entrypoints - I can select them from a list but certain ones don't do anything or display the same as another
In the connections menu you mentioned previously (I was looking for filepaths) i tried to specify the other entrypoint that wasn't mentioned and it said it was wrong and offered to "fix"

[chr0nzz]

Found the issue - there was a misconfigured proxy header setting (x_prefix) that caused Traefik to mangle all URL routing when it sent X-Forwarded-Prefix. Just pushed a fix, pull the latest :beta and the domain should work again alongside the entrypoint middleware display.

Both of those are expected behavior:

The "fix" prompt - that's the self-route entrypoint validator in Settings → Connection → Self Route. TM checks the entrypoint you enter against the live Traefik API. If it can't find a match it warns you and suggests the closest one. If your entrypoint is internal-only and not showing in the Traefik API, you can dismiss the warning - it won't affect routing.

Entrypoints showing the same - if multiple entrypoints have the same middleware defined in your traefik.yml, the ep chips will be identical on all matching routes. That's correct - they inherit the same middleware.

"don't do anything" - if an entrypoint isn't returned by your Traefik API it won't appear in the chip selector, or if selected it won't be saved. Make sure your Traefik API is reachable from TM and returning all entrypoints at /api/entrypoints.

[Xanderlicious]

I'm back connected using the domain

I've also had a bit of a closer look and whilst the route map has changed (quite substancially to how it was before) it actually isn't any different - there are still services that are not reflecting the middleware thats applied at the entrypoint.

And with regards to the whole entrypoint issue - let me try and explain with pictures

When I first select route map, I get this view:

If I select websecure-ext - I get this view

If I select websecure-int.....

and then if I go back to websecure-ext....

Its exactly the same and my websecure-ext routes/services are not showing at all

[chr0nzz]

Just pushed another fix - TM now falls back to the live Traefik API to resolve entrypoints for routes that don't have them explicitly defined in the config file. Pull the latest :beta and the Route Map filter and ep chips should work correctly for all your routes.

[Xanderlicious]

Further debugging shows _build_all_apps() is returning correct entrypoints for all routes — no missing entries. The issue appears to be in the Route Map frontend filter rather than the backend data. All 50 routes have entrypoints correctly assigned.
Confirmed via:

docker exec traefik-manager python3 -c "
import sys
sys.path.insert(0, '/app')
from app import _build_all_apps
apps, _ = _build_all_apps(include_external=True, include_internal=False)
for a in apps:
eps = a.get('entryPoints', [])
if not eps:
print('MISSING: ' + a['name'])
else:
print('OK: ' + a['name'] + ' -> ' + str(eps))
"

All routes returned OK with correct entrypoints. The Route Map filter shows the same routes regardless of which entrypoint chip is selected (websecure-ext and websecure-int show identical results). This points to a bug in the Route Map JS filter logic rather than the Python backend.

this also might explain why when selecting routes in the android app, it crashes the app altogether?

[chr0nzz]

Thanks for the detailed report. Found the bug: if any route in your Traefik config uses a YAML scalar for entryPoints (e.g. entryPoints: websecure instead of entryPoints: [websecure]) the backend was passing it as a string. JavaScript's String.prototype.includes is a substring search, so a route with entryPoints: "websecure" would incorrectly match a filter for "web". This is fixed in the next release - the backend now always normalizes to a list, and the Route Map uses an array-safe helper everywhere.

That said, for your specific websecure-ext/websecure-int case, neither name is a substring of the other so the scalar bug wouldn't produce identical results. The most likely explanation there is that all your routes genuinely have both entrypoints assigned. You can now verify this using the Entry Point chips added to each route's popup in the Route Map.

For the Android crash: if it happens when opening a route's detail screen, the scalar entryPoints fix closes that path (calling .join() on a string throws a TypeError). If the crash persists after the update, please share a logcat stack trace so I can pinpoint it exactly.