Cloud Deployment Skill

Manage complete cloud deployment lifecycle from local validation to production release, including infrastructure provisioning, health monitoring, and rollback procedures.

When to Use This Skill

Deploy application to staging or production
Provision cloud infrastructure
Update deployed application
Monitor application health
Troubleshoot deployment issues
Rollback problematic deployments

What This Skill Does

Local Validation: Verify tests, linting, builds pass locally
Infrastructure Check: Ensure cloud resources provisioned correctly
Staging Deployment: Deploy to staging environment
Smoke Testing: Run critical path tests on staging
Production Deployment: Deploy to production with approval
Health Monitoring: Track metrics, errors, performance
Rollback: Revert to previous version if issues detected

Step-by-Step Procedure

Phase 1: Pre-Deployment Validation (Local)

1. Run full test suite
   npm test || php artisan test || pytest

2. Run linting
   npm run lint || phpstan analyze

3. Run type checking
   tsc --noEmit || phpstan --level=8

4. Build production assets
   npm run build || php artisan build

5. Verify build artifacts created

SUCCESS CRITERIA:
- [ ] All tests pass
- [ ] No lint errors
- [ ] No type errors
- [ ] Build completes
- [ ] Artifacts valid

Phase 2: Infrastructure Verification

1. Check cloud configuration files exist:
   - .env.production
   - docker-compose.yml (or equivalent)
   - Infrastructure as Code (Terraform/ARM)

2. Verify required services:
   - Database accessible
   - Cache server running
   - Queue workers ready
   - Storage buckets exist

3. Check security:
   - Secrets in vault (not code)
   - API keys rotated
   - SSL certificates valid
   - Firewall rules correct

4. Verify backup procedures:
   - Database backup recent
   - Rollback plan documented
   - Backup restoration tested

Phase 3: Staging Deployment

1. Create deployment branch:
   git checkout -b deploy/staging-$(date +%Y%m%d%H%M%S)

2. Trigger staging deployment:
   - Via GitHub Actions workflow
   - Via cloud provider CLI
   - Via infrastructure tool

3. Wait for deployment completion (max 5 min)

4. Verify deployment status:
   - Pods/containers running
   - Health endpoints responding
   - Database migrations applied
   - Static assets uploaded

Phase 4: Staging Validation

SMOKE TESTS (critical paths only):
1. Health check: GET /health → 200 OK
2. Auth flow: Login → Success
3. Core API: GET /api/resource → 200 OK
4. Database: Read/write test → Success

MONITORING:
1. Check error rates (should be 0)
2. Check response times (<500ms)
3. Review application logs
4. Verify no crash loops

WAIT TIME: 3-5 minutes for system stabilization

Phase 5: Production Readiness

1. Compare staging vs production configs
   - Environment variables
   - Resource limits
   - Scaling settings
   - Feature flags

2. Review deployment logs for warnings

3. Verify rollback ready:
   - Previous version tagged
   - Rollback procedure documented
   - Backup confirmed recent

4. Get approval (if required):
   - Manual approval gate
   - Stakeholder sign-off
   - Change request ticket

Phase 6: Production Deployment

1. Create deployment PR (if required)
   - Title: "Deploy v1.2.3 to production"
   - Description: Changes since last deploy
   - Link to staging validation results

2. Trigger production deployment:
   - Via GitHub Actions
   - Via cloud provider
   - Via infrastructure tool

3. Wait for deployment (max 10 min)

4. Monitor deployment progress:
   - Rolling update status
   - Pod/container replacements
   - Zero-downtime verification

Phase 7: Post-Deployment Monitoring

IMMEDIATE CHECKS (first 5 minutes):
1. Health endpoints → 200 OK
2. Error rate → <0.1%
3. Response time → baseline ±10%
4. No crash loops
5. Database connections stable

EXTENDED MONITORING (next 30 minutes):
1. User-facing features working
2. Background jobs processing
3. No unusual traffic patterns
4. No security alerts
5. Performance within SLOs

ALERT THRESHOLDS:
- Error rate >1% → Investigate
- Response time >2x baseline → Investigate
- Health check fails → Rollback
- Critical feature broken → Rollback

Phase 8: Rollback (If Needed)

TRIGGERS:
- Health checks failing
- Error rate >5%
- Critical feature broken
- Data corruption detected
- Security vulnerability

PROCEDURE:
1. Alert team via Slack/Discord
2. Revert to previous deployment:
   - Tag previous version
   - Trigger rollback workflow
   - Or manually: kubectl rollout undo

3. Verify rollback success:
   - Health checks pass
   - Error rate normal
   - Services stable

4. Create incident issue:
   - What failed
   - When detected
   - Rollback actions taken
   - Root cause analysis needed

Expected Input

User Request Examples:

"Deploy to staging"
"Release v1.2.3 to production"
"Check production health"
"Rollback last deployment"

Deployment Configuration:

# .github/workflows/deploy-production.yml
name: Deploy to Production
on:
  workflow_dispatch:
    inputs:
      version:
        description: 'Version to deploy'
        required: true

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      - name: Deploy to Cloud
        run: |
          # Cloud provider specific deployment

Expected Output

Deployment Report:

## Production Deployment Complete ✅

**Version**: v1.2.3
**Environment**: Production
**Deploy Time**: 2026-01-12 14:30 UTC
**Duration**: 8 minutes

### Pre-Deployment
- [x] Tests passed (120/120)
- [x] Linting passed
- [x] Build completed
- [x] Infrastructure verified

### Staging
- [x] Deployed successfully
- [x] Smoke tests passed (5/5)
- [x] Performance baseline met
- [x] No errors in logs

### Production
- [x] Deployed successfully
- [x] Zero downtime achieved
- [x] Health checks passing
- [x] Error rate: 0.02% (normal)
- [x] Response time: 245ms avg (baseline: 250ms)

### Monitoring
- Error rate: 0.02%
- Response time: 245ms (p50), 480ms (p95)
- Uptime: 100%
- Traffic: 1,200 req/min (normal)

### Changes Deployed
- Feature: User authentication (#42)
- Fix: Password reset email (#43)
- Refactor: Optimize database queries (#44)

**Next deployment window**: 2026-01-15 14:00 UTC

Cloud Provider Commands

Azure

# Deploy to App Service
az webapp deployment source config-zip \
  --resource-group myResourceGroup \
  --name myApp \
  --src ./build.zip

# Check health
az webapp show --name myApp --resource-group myResourceGroup

AWS

# Deploy to Elastic Beanstalk
eb deploy production

# Check health
eb health production

GCP

# Deploy to Cloud Run
gcloud run deploy myapp \
  --image gcr.io/project/myapp:v1.2.3 \
  --region us-central1

# Check health
gcloud run services describe myapp --region us-central1

Kubernetes

# Deploy
kubectl apply -f k8s/production/

# Check rollout
kubectl rollout status deployment/myapp

# Rollback
kubectl rollout undo deployment/myapp

Integration with Other Skills

Execution: Called by task-execution
Testing: Requires test-generation to pass
Monitoring: Uses health-monitoring

Configuration

Deployment Environments:

Development: Auto-deploy on main branch push
Staging: Manual trigger or scheduled
Production: Manual approval required

Health Check Endpoints:

GET /health → 200 OK
GET /health/database → 200 OK
GET /health/cache → 200 OK
GET /health/queue → 200 OK

Rollback Criteria:

Auto-rollback if:
- Health check fails for >2 minutes
- Error rate >10%
- Critical endpoint returns 5xx

Manual rollback if:
- Data corruption detected
- Security issue found
- Business-critical feature broken

References

Cloud provider documentation
.github/workflows/deploy-*.yml - Deployment workflows
docs/DEPLOYMENT.md - Deployment procedures
Infrastructure as Code files (Terraform, ARM, CloudFormation)

cloud-deployment