Issue Description

Our Node.js application deployed on Railway (Asia Southeast region: asia-southeast1-eqsg3a) is experiencing constant ETIMEDOUT errors when attempting to connect to Meta's WhatsApp Cloud API at graph.facebook.com. The same code works perfectly in local development and was working in production until recently.

Impact

• Severity: High - Core application functionality broken

• Frequency: Constant failures (100% error rate currently)

• Affected Service: Backend API service

• Region: railway/asia-southeast1-eqsg3a (visible in response headers: X-Railway-Edge: railway/asia-southeast1-eqsg3a)

Environment Details

• Platform: Railway

• Region: Asia Southeast 1

• Runtime: Node.js 22.21.0

• Deployment: Successful build, fails at runtime

• Target Host: graph.facebook.com (Meta/Facebook Graph API)

• Target URL: https://graph.facebook.com/v21.0/{business-account-id}/message_templates

Error Details

Error Code: ETIMEDOUT

Error Type: TypeError: fetch failed with AggregateError cause

Description: TCP connection timeout when attempting to reach Meta's API servers

Complete Error Log

{
  "timestamp": "2025-10-25 14:27:06",
  "error": "Error fetching templates fetch failed",
  "name": "TypeError",
  "cause": {
    "code": "ETIMEDOUT"
  },
  "causeName": "AggregateError",
  "causeCode": "ETIMEDOUT",
  "url": "https://graph.facebook.com/v21.0/123123123/message_templates",
  "businessAccountId": "123123123",
  "hasAccessToken": true,
  "stack": "TypeError: fetch failed\n    at node:internal/deps/undici/undici:14900:13\n    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)\n    at async Object.listTemplates (file:///app/server/dist/lib/message-provider.js:157:32)\n    at async MessageService.listTemplates (file:///app/server/dist/services/message.service.js:508:27)\n    at async file:///app/server/dist/routes/messages.js:146:27"
}

Deployment Logs (Successful Build)

2025-10-25T10:35:47.349825076Z [inf]  
> @event-management/server@1.0.0 build /app/server
> tsc --build --force && tsc-alias

2025-10-25T10:35:54.396484170Z [inf]  copy /app/node_modules
2025-10-25T10:35:59.291071711Z [inf]  [92mBuild time: 134.90 seconds[0m

Build completes successfully. The application starts without errors.

Runtime Logs (Connection Failures)

2025-10-25 14:06:18:618 [error]: Error occurred:
{
  "error": "fetch failed",
  "stack": "Error: fetch failed\n    at Object.listTemplates (file:///app/server/dist/lib/message-provider.js:232:23)\n    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)\n    at async MessageService.listTemplates (file:///app/server/dist/services/message.service.js:508:27)\n    at async file:///app/server/dist/routes/messages.js:146:27",
  "path": "/messages/templates",
  "method": "GET",
  "user": "1321321323"
}

2025-10-25 14:27:06:276 [error]: Error fetching templates fetch failed
{
  "name": "TypeError",
  "cause": {
    "code": "ETIMEDOUT"
  },
  "causeName": "AggregateError",
  "causeCode": "ETIMEDOUT",
  "url": "https://graph.facebook.com/v21.0/123123123/message_templates",
  "businessAccountId": "123123123",
  "hasAccessToken": true
}

What We've Ruled Out

Environment variables: Confirmed all credentials are correctly configured

Code issues: Same code works in local development (returns 200 OK)

Authentication: hasAccessToken: true - credentials are present

Rate limiting: No 429 errors, well within Meta's API limits

Application timeout: Using 30-second AbortController timeout, but TCP connection times out before that

Network Connectivity Test Results

• Local Development (macOS):

  • fetch('https://graph.facebook.com/...') → 200 OK

  • Response time: ~300ms

  • Works consistently

• Railway Production (asia-southeast1):

  • fetch('https://graph.facebook.com/...') → ETIMEDOUT

  • Connection times out before establishing TCP connection

  • Fails consistently (was intermittent, now constant)

Timeline

• Previously: Application worked correctly in production

• Recently: Started experiencing intermittent failures

• Currently: 100% failure rate - all requests timeout

Expected Behavior

Railway containers should be able to establish TCP connections to graph.facebook.com (Meta's public API) within reasonable timeout periods (~5-10 seconds).

Request for Support

Could you please investigate:

1. Network routing from asia-southeast1 region to Meta's API servers

2. DNS resolution for graph.facebook.com from Railway containers

3. Firewall rules that might be blocking connections to Meta/Facebook IP ranges

4. Regional network issues specific to asia-southeast1-eqsg3a

Could this be a Railway infrastructure/network issue as the same application code works in other environments?