Title: Healthcheck fails on every deploy — app starts cleanly, PORT confirmed correct, but zero HTTP traffic ever reaches the container

Service is failing healthcheck on 5 consecutive deployments, including a manual redeploy of unchanged code and one after explicitly pinning PORT to match the domain's target port. Build and Deploy stages succeed; only Network > Healthcheck fails.

Deployment IDs:

bc1db2a5-aaa7-4ed9-afdc-9eb500f7c603

a7681155-d5e0-4865-bfd3-3948ecbfce62

00e9d59c-9351-4aae-87a6-174cd2498999

40d51187-0fed-421c-9beb-0cfcf0a41e18

be98b62b-3477-4294-8024-eb5cd6156378 (after pinning PORT=8080 to match target port — still failed)

What I see in logs (railway logs --deployment ):

▲ Next.js 14.2.35

• Local: http://localhost:8080

✓ Starting...

✓ Ready in ~210-285ms

Stopping Container

App starts and is ready in ~250ms every time. Container is killed shortly after.

railway logs --http --deployment returns zero HTTP log lines for every failed deployment — not even the healthcheck probe request itself shows up.

Ruled out so far:

PORT/target-port mismatch (explicitly pinned PORT=8080, matches domain's target port — still fails)

App-level error (clean startup logs, no errors)

Platform-wide incident (status.railway.com reports UP)

The currently-live deployment (same service, deployed 2026-08-25) is healthy right now and correctly serving traffic on the same port/path, so this isn't a config or app-level issue on my end.

Project: Sales-Forecast-sergio, Service: forecast-sales-ai, Environment: production, Region: cdg1 (Paris).

Any idea what could cause the healthcheck probe to never reach a newly-deployed container while the app itself starts fine and listens on the correct port?