Skip to content

deployment_checklist.md

Latest commit

 

History

History
198 lines (159 loc) · 5.72 KB

deployment_checklist.md

File metadata and controls

198 lines (159 loc) · 5.72 KB
Raw

Production Deployment Checklist

Pre-Deployment Preparation

✅ Code Cleanup

  • Removed test files (quick_firestore_test.py, test_query.json, validate_firestore_setup.py)
  • Removed old query history data (query_history.json)
  • Updated all imports to use cloud-ready components
  • Added proper error handling and logging
  • Verified no development dependencies in production requirements

✅ Configuration

  • Environment variables properly set in .env
  • Google Cloud project configured
  • Firestore database created and operational
  • BigQuery datasets and tables verified
  • API endpoints tested and working
  • CORS configured for production domains

✅ Security

  • Sensitive data in environment variables (not hardcoded)
  • Service account permissions minimal and appropriate
  • API input validation implemented
  • Error messages sanitized (no sensitive info leaked)
  • HTTPS enforced (Cloud Run default)

Cloud Infrastructure Setup

Google Cloud Project Setup

# Set up the project
gcloud config set project gen-lang-client-0044046698

# Enable required APIs
gcloud services enable run.googleapis.com
gcloud services enable bigquery.googleapis.com
gcloud services enable firestore.googleapis.com
gcloud services enable aiplatform.googleapis.com

Firestore Database

  • Database created in us-central1 region
  • Native mode selected
  • Security rules configured (if needed)
  • Collections properly indexed

BigQuery Setup

  • Dataset: home_buyer_hackathon_data
  • Tables: listings, neighborhoods, affordability_parameters
  • Proper IAM permissions for Cloud Run service account
  • Data loaded and verified

Deployment Steps

1. Container Preparation

# Build the container image
docker build -t gcr.io/gen-lang-client-0044046698/adk-home-buyer .

# Test locally
docker run -p 8000:8000 --env-file .env gcr.io/gen-lang-client-0044046698/adk-home-buyer

# Push to Google Container Registry
docker push gcr.io/gen-lang-client-0044046698/adk-home-buyer

2. Cloud Run Deployment

# Deploy to Cloud Run
gcloud run deploy adk-home-buyer \
  --image gcr.io/gen-lang-client-0044046698/adk-home-buyer \
  --platform managed \
  --region us-central1 \
  --allow-unauthenticated \
  --set-env-vars GOOGLE_CLOUD_PROJECT=gen-lang-client-0044046698 \
  --set-env-vars QUERY_HISTORY_BACKEND=firestore \
  --set-env-vars ENABLE_QUERY_HISTORY=true \
  --memory 2Gi \
  --cpu 2 \
  --max-instances 10 \
  --port 8000

3. Frontend Deployment

# Build React frontend
cd frontend
npm run build

# Deploy to Firebase Hosting, Netlify, or other static host
# Update API_BASE_URL to point to Cloud Run service

Post-Deployment Verification

✅ Health Checks

  • API health endpoint responding: GET /api/health
  • Backend status endpoint working: GET /api/history/status
  • Sample analysis request successful: POST /api/analyze
  • Firestore query history being saved
  • BigQuery queries executing successfully
  • Vertex AI models responding

✅ Performance Testing

  • Load testing with multiple concurrent requests
  • Response times under acceptable thresholds
  • Auto-scaling working properly
  • Memory and CPU usage within limits

✅ Monitoring Setup

  • Cloud Logging configured and working
  • Error alerting set up
  • Performance monitoring enabled
  • Uptime checks configured

Environment Variables for Production

Required Environment Variables

# Core Configuration
GOOGLE_CLOUD_PROJECT=gen-lang-client-0044046698
PORT=8000

# Query History
QUERY_HISTORY_BACKEND=firestore
FIRESTORE_COLLECTION=query_history
ENABLE_QUERY_HISTORY=true
MAX_HISTORY_ENTRIES=50

# BigQuery Configuration
BIGQUERY_DATASET=home_buyer_hackathon_data
BIGQUERY_LOCATION=northamerica-northeast2
LISTINGS_TABLE=listings
NEIGHBORHOODS_TABLE=neighborhoods

# Vertex AI Configuration
VERTEX_AI_LOCATION=us-central1
DEFAULT_AGENT_MODEL=gemini-2.0-flash-001
ORCHESTRATOR_MODEL=gemini-2.0-flash-001
EMBEDDING_MODEL_NAME=text-embedding-004

# Feature Flags
USE_MOCK_DATA=false

Rollback Plan

In case of deployment issues:

  1. Immediate Rollback: Use Cloud Run revisions to rollback to previous version

    gcloud run services update-traffic adk-home-buyer --to-revisions=PREVIOUS_REVISION=100

  2. Database Issues: Firestore has automatic backups, BigQuery has versioned tables

  3. Configuration Issues: Update environment variables without redeployment

    gcloud run services update adk-home-buyer --set-env-vars KEY=VALUE

Maintenance Procedures

Regular Tasks

  • Monitor Cloud Run logs for errors
  • Review Firestore usage and clean up old queries if needed
  • Monitor BigQuery costs and optimize queries
  • Update dependencies monthly
  • Review and rotate service account keys (if applicable)

Scaling Considerations

  • Monitor request patterns and adjust instance limits
  • Consider regional deployment for global users
  • Optimize database queries for performance
  • Implement caching if needed

Success Criteria

✅ Deployment is successful when:

  • API responds to health checks
  • Frontend can connect to API
  • Property searches return results
  • Query history is being saved to Firestore
  • No errors in Cloud Run logs
  • Response times < 10 seconds for analysis requests
  • System handles multiple concurrent users

Support Contacts

  • Development Team: [Your team contact]
  • Google Cloud Support: [Your support plan]
  • Incident Response: [Your on-call procedures]

Deployment Date: ________________
Deployed By: ___________________
Version: ______________________
Rollback Plan Tested: ✅ / ❌