Railway Production Deployment

Problem

Deploying to Railway requires coordinating multiple steps: environment variables, database migrations, health checks, and external service configuration (Stripe, SendGrid, GitHub OAuth). Missing any step can cause production failures.

Trigger Conditions

Use this skill when:

•User asks to "deploy to production" or "deploy to Railway"
•Completing a feature and ready for production release
•Initial production launch setup
•Updating production after bug fixes

Workflow

Phase 1: Pre-Deployment Validation (30 min)

Run locally before deploying:

bash

# 1. Validate environment variables
npm run validate:env

# 2. Run all tests
npm test
# Expected: ✓ 507 passed

# 3. Lint and type-check
npm run lint
npx tsc --noEmit

# 4. Build verification
npm run build

# 5. Verify clean working directory
git status
git log -5

Stop if any validation fails. Fix issues before proceeding.

Phase 2: Railway Configuration

Critical environment variables to set in Railway dashboard:

Variable	Source	Example
`DATABASE_URL`	Auto-set by Railway Postgres	-
`REDIS_URL`	Auto-set by Railway Redis	-
`AUTH_SECRET`	`openssl rand -base64 32`	-
`STRIPE_SECRET_KEY`	Stripe Dashboard (live mode)	`sk_live_...`
`NEXT_PUBLIC_STRIPE_PUBLISHABLE_KEY`	Stripe Dashboard	`pk_live_...`
`STRIPE_WEBHOOK_SECRET`	Stripe webhook config	`whsec_...`
`CRON_SECRET`	Generate random	-
`HONEYBADGER_API_KEY`	Honeybadger dashboard	-

See PRODUCTION_DEPLOYMENT.md for full list (30+ variables).

External service updates:

•Stripe: Configure webhook https://codesalvage.com/api/webhooks/stripe
•GitHub OAuth: Update callback URL to production domain
•Configure custom domain in Railway → Settings → Domains

Phase 3: Deploy

bash

# Auto-deploy on push to main
git push origin main

# OR manual deploy via Railway CLI
railway up

# Monitor deployment
railway logs

Phase 4: Database Migrations

bash

# CRITICAL: Run after deployment succeeds
railway run npm run db:migrate:deploy

Phase 5: Post-Deployment Verification (15 min)

bash

# 1. Automated health check
bash scripts/post-deployment-check.sh https://codesalvage.com YOUR_CRON_SECRET

# 2. Detailed health check
curl -H "Authorization: Bearer YOUR_CRON_SECRET" \
  "https://codesalvage.com/api/health?detailed=true" | jq

# 3. Manual checks
# - Homepage loads: https://codesalvage.com
# - GitHub OAuth login works
# - Browse projects: /projects
# - Health endpoint: /api/health returns 200

Phase 6: Monitoring (First 24-48 Hours)

Check every 2 hours:

•Honeybadger (no critical errors)
•Railway metrics (CPU, memory normal)
•Stripe Dashboard (webhooks delivering)
•SendGrid Activity (emails sending)

Success Criteria

Deployment successful when:

•✅ All health checks pass
•✅ No errors in Honeybadger
•✅ Railway metrics normal
•✅ Stripe webhooks delivering
•✅ DNS resolving correctly
•✅ HTTPS certificate valid

Rollback Plan

If critical issues found:

bash

# In Railway dashboard:
# Deployments → Click previous successful deployment → Redeploy

Then investigate with railway logs and Honeybadger.

Common Issues

Symptom	Fix
Database connection errors	Check `DATABASE_URL`, verify Postgres running
Redis connection errors	Check `REDIS_URL`, verify Redis running
Stripe webhook failures	Verify webhook URL and signing secret in Stripe Dashboard
Email not sending	Check SendGrid API key, verify domain authentication
GitHub OAuth fails	Verify callback URL: `https://codesalvage.com/api/auth/callback/github`

References

•Full guide: DEPLOYMENT_WORKFLOW.md
•Detailed deployment: PRODUCTION_DEPLOYMENT.md
•Smoke testing: tests/SMOKE_TESTING_CHECKLIST.md
•Load testing: tests/load-testing/README.md

Example

Typical deployment session:

bash

# Local validation
npm run validate:env && npm test && npm run lint && npm run build

# Deploy
git push origin main

# Monitor
railway logs

# Migrate
railway run npm run db:migrate:deploy

# Verify
bash scripts/post-deployment-check.sh https://codesalvage.com $CRON_SECRET

# Success! Monitor Honeybadger and Railway metrics