the most common reason for this issue is a dns misconfiguration at the apex/root level

two things to check:

first, if you're on cloudflare, make sure the proxy is turned off (set to dns only, grey cloud) and your ssl/tls mode is set to "full" not "full strict"

second, make sure your app is listening on railway's PORT environment variable and not a hardcoded port like 3000 or 8080, railway injects the port at runtime so if it's hardcoded it won't work even if dns is fine

hope this help you, if not which dns provider are you using? that would help narrow it down further

namecheap

for namecheap, set up a cname record on the root domain (@) pointing to your railway-provided domain (something like abc123.up.railway.app)

can you check in your namecheap dns settings if that cname record on @ is actually there and pointing to the correct railway value? also make sure there are no conflicting a records or redirect records set for the same @ host, namecheap warns those can conflict

It is. It is pointing to the correct railway provided domain. There is another platform using the @ however that is a TXT record and not a CNAME.

ok can you share what records you currently have on the @ host in namecheap ?

Brevo for emails.

It's just that. And then I have setup the rest for Railway.

One for the main domain and one for the backend(which works. I am able to reach the domain successfully and get my docs)

However the main domain pointing to my web application throws the 502.

so the dns is likely fine then since the backend works. this points to an application-level issue not dns, is your frontend app listening on the PORT environment variable that railway provides, or is it hardcoded to a specific port?

• Port: Not hardcoded. It comes from the PORT environment variable.
• Default:80 when PORT is not set.
• Flow:
  1. docker-entrypoint.sh sets PORT="${PORT:-80}" (default 80).
  2. nginx.conf.template uses listen ${PORT};.
  3. The entrypoint runs sed to replace ${PORT} in the template before starting nginx.

Local dev (Vite)

• Vite dev server typically uses port 5173 (its default).

Summary

FilePurposenginx.conf.templatelisten ${PORT}; — configurable via envnginx.conflisten 80; — hardcoded for local usedocker-entrypoint.shPORT="${PORT:-80}"

And it's the same port 80 that I have configured on Railway.

This resolved it. It was supposed to be listening to 8080 but it was set to 80 on Railway

Thank you very much!