Hi Railway Team,

I need help. I've been on Railway Pro for 13 days ($20/mo), but my Node.js service is completely non-functional due to a platform-level memory limitation.

## Problem

Node.js heap is hard-capped at ~14 MB, preventing my production API from operating.

- Current heap: 14.10 MB
- Required: 512 MB minimum
- Status: Circuit breaker blocking all traffic at 83% memory
- Downtime: 13 days (100% of Pro subscription)
- Service: https://specular-production.up.railway.app

## Impact

- 49 users blocked
- $32,000 in user capital inaccessible
- Paying $20/mo for 0% uptime

## What I've Tried (All Failed)

I've systematically tried every configuration method over 13 days. All failed:

**1. Environment Variable (March 1)**

NODE_OPTIONS=--max-old-space-size=512

Result: Ignored. Heap: 13.86 MB

**2. package.json start script**
```json
"start": "node --max-old-space-size=512 src/api/MultiNetworkAPI.js"

Result: Ignored. Heap: 14.11 MB

3. railway.json startCommand

"startCommand": "NODE_OPTIONS='--max-old-space-size=512' npm start"

Result: Ignored. Heap: 14.36 MB

4. nixpacks.toml

[variables]
NODE_OPTIONS = "--max-old-space-size=512"

Result: Ignored. Heap: 14.61 MB

5. Custom Dockerfile (TODAY)

CMD ["node", "--max-old-space-size=512", "src/api/MultiNetworkAPI.js"]

"build": {"builder": "DOCKERFILE"}

Result: STILL IGNORED. Heap: 14.10 MB

Evidence: https://specular-production.up.railway.app/stats

What This Proves

Even with a custom Dockerfile where I have full control over the Node.js start command, Railway is overriding the heap allocation at the platform level.

This is NOT a configuration issue. Railway is enforcing a ~14 MB limit that ignores:

• Environment variables
• Start command flags
• Dockerfile CMD directives

Previous Support Attempt

March 1: Railway suggested NODE_OPTIONS=--max-old-space-size=512

• Already tried (was our first attempt)
• Still failed
• No follow-up response for 12 days

What I Need

Option 1 (Preferred): Engineering Investigation - 48 Hours

Urgent review of:

1. Why is Railway overriding Dockerfile CMD memory flags?
2. Is there a platform limitation on Node.js heap?
3. Are Pro resources properly provisioned?

Option 2: Full Refund + Migration

If unfixable within 48 hours:

• Refund for 13 days of non-working service
• I'll migrate to alternative platform

Why This Is Urgent

After 13 days:

• ✅ Tried all 5 configuration methods
• ✅ Provided custom Dockerfile with explicit control
• ✅ Proven it's a platform issue
• ✅ Paying for 0% uptime

I cannot wait longer. Users are blocked and I'm paying for a broken service.

Technical Details

Test locally (same Dockerfile):

docker build -t specular .
docker run specular

Result: 512 MB heap allocated ✅

Railway deployment:

Result: 14 MB heap ❌

Conclusion: Railway is overriding Docker CMD at runtime.

Next Steps

If no substantive response within 48 hours, I will:

1. Request chargeback
2. Migrate to Render/Fly.io
3. Document publicly

I prefer to stay with Railway, but need this resolved urgently.

Project: Specular

GitHub: github.com/thegrand-canyon/specular

Live endpoint: https://specular-production.up.railway.app/stats

Available for immediate debugging session with your team.