Description of the issue:

My service (whatsapp-bot-v3) has had 3 consecutive deployment attempts fail at Network → Healthcheck — 2 fresh builds from new commits, and 1 plain "Redeploy" of an already-built image (no rebuild). All three failed identically, while the previous deployment from ~1 day ago remains ACTIVE and healthy.

Deploy logs show the app starting cleanly every time:

• PostgreSQL connection established
• Server running on port 3000
• Health: http://localhost:3000/
• Webhook: http://localhost:3000/webhook
• Scheduler starts and ticks normally every minute for 4+ minutes (morning routes, attendance reminders, tracking link reminders all logging successfully)

Despite this, Railway's healthcheck against "/" reports "service unavailable" on every one of 14 retry attempts over ~5 minutes, and the deployment is marked failed.

What I've tried:

• Restarted the previous ACTIVE deployment — worked fine, no issue.
• Redeployed the failing image directly (no rebuild) — still failed healthcheck.
• Pushed a trivial no-op code change to force a fresh build — still failed healthcheck, identical pattern.

This strongly suggests a networking/infra-side issue between the healthcheck prober and this specific service/deployment, not an application-level problem, since the app is demonstrably alive and serving internally.

railway.json:

{

"build": { "builder": "NIXPACKS", "buildCommand": "npm install && npm run build" },

"deploy": { "startCommand": "node dist/index.js", "healthcheckPath": "/" }

}

Error messages: "Deployment failed during the network process" / "Healthcheck failed" / "1/1 replicas never became healthy!"