Just a quick follow-up to my previous message:

In my app, the URLs are correctly defined — I’m clicking on links like https://myapp.com/... and there’s no mention of port 8080 anywhere in my code or config.

However, when inspecting the network traffic (via browser dev tools), I notice that some of these requests hang for 50–60 seconds, and then result in a redirect to https://myapp.com:8080/..., which fails to load.

This only happens on certain URLs and at unpredictable times, making it very hard to reproduce consistently, but it’s clearly coming from the platform’s behavior, not the application code.

Thanks again for your help,

Hmmm, on your apache config, can you try adding:

```
RemoteIPHeader X-Forwarded-For

RemoteIPInternalProxy 0.0.0.0/0

SetEnvIf X-Forwarded-Proto "https" HTTPS=on
```

Hey,

0.0.0.0/0 is not possible (Apache error), so I put this, but still doesn't work

```

# Enable RemoteIP mod

RUN a2enmod remoteip

# Config mode

RUN echo 'RemoteIPHeader X-Forwarded-For\n\

RemoteIPInternalProxy 10.0.0.0/8\n\

RemoteIPInternalProxy 172.16.0.0/12\n\

SetEnvIf X-Forwarded-Proto "https" HTTPS=on' \

> /etc/apache2/conf-available/remoteip.conf \

&& a2enconf remoteip

```

Same issue. Do you have any other idea?

Hey kaleidoscopique, picking up for Angelo.

Definitely something is happening where Apache is not handling traffic properly in some scenarios when coming through Railway's proxy.

First: Can you share your full apache config and Dockerfile? That will help me debug.

I think I want to look at X-Forwarded-Proto, X-Forwarded-Host, and X-Forwarded-Port headers first, then look at ensuring we're handling all the RemoteIPInternalProxy proxy headers correctly.

Other ideas:

```
# Prevent port from leaking into redirects

RequestHeader set X-Forwarded-Port "443"

RequestHeader set X-Forwarded-Proto "https"

# Use X-Forwarded-Host for ServerName

UseCanonicalName Off
```

Hello! Thanks a lot for your help.

I’ve restructured my deployment to remove Apache entirely, switching to a simple setup using the built-in PHP CLI server (php -S 0.0.0.0:8080 -t public).

Since making this change:

• All redirects are now working properly.

• There are no more unwanted :8080 ports leaking into URLs.

• The app behaves consistently, with no proxy-related issues.

In the end, I realized I didn’t actually need Apache for my project’s requirements, and using PHP’s native server fits my Railway deployment much more smoothly.

But FYI, I can share my old entrypoint : https://pastebin.com/B4ZgiSMC

and my old dockerfile : https://pastebin.com/4Ma3GUhJ

Nothing fancy, but if you have any idea about why it didn't work, I'm really curious.

Thanks again! Have a nice day.

I think you were seeing routing issues with how Apache was configured. Glad to hear you were able to work around the issue!