Hands-onLesson 10 of 12

Real-world Scenarios: Hands-on Labs

Deploy production-grade apps. Practice multi-tier deployments, scaling, monitoring, and cost optimization.

Lab 1: Deploy Web App with Auto-Scaling

bash
# Create instance template
gcloud compute instance-templates create web-app-template \
  --machine-type=n1-standard-1 \
  --image-family=debian-11 \
  --image-project=debian-cloud \
  --metadata=startup-script='#!/bin/bash
apt-get update && apt-get install -y nodejs npm
cat > /app.js << EOF
const http = require("http");
http.createServer((req, res) => {
  res.end("Hello from GCP! " + new Date());
}).listen(8080);
EOF
node /app.js &'

# Create managed instance group
gcloud compute instance-groups managed create web-app-ig \
  --base-instance-name=web --template=web-app-template \
  --size=3 --zone=us-central1-a

# Health check
gcloud compute health-checks create http web-health \
  --global --port=8080

# Auto-scaling
gcloud compute instance-groups managed set-autoscaling web-app-ig \
  --max-num-replicas=10 --min-num-replicas=3 \
  --target-cpu-utilization=0.7 --zone=us-central1-a

# Load balancer
gcloud compute backend-services create web-backend \
  --global --protocol=HTTP --health-checks=web-health
gcloud compute backend-services add-backend web-backend --global \
  --instance-group=web-app-ig --instance-group-zone=us-central1-a
gcloud compute url-maps create web-lb --default-service=web-backend
gcloud compute target-http-proxies create web-proxy --url-map=web-lb
gcloud compute forwarding-rules create web-forwarding-rule \
  --global --target-http-proxy=web-proxy --ports=80

Lab 2: Add Database Layer

bash
# Cloud SQL instance
gcloud sql instances create web-app-db \
  --database-version=POSTGRES_13 --tier=db-f1-micro \
  --region=us-central1 --availability-type=REGIONAL

# Create DB and user
gcloud sql databases create app_db --instance=web-app-db
gcloud sql users create app_user --instance=web-app-db \
  --password=SecurePass123

# Enable private IP
gcloud sql instances patch web-app-db --assign-ip

# Connect from app via Cloud SQL Auth Proxy
# gcloud cloud-sql-proxy -instances=PROJECT:us-central1:web-app-db=tcp:5432

Lab 3: Object Storage for Media

bash
# Create bucket
gsutil mb -l us-central1 gs://my-app-media

# Upload file
gsutil cp photo.jpg gs://my-app-media/

# Generate signed URL (1 hour)
gsutil signurl ~/.ssh/key.json -d 1h gs://my-app-media/photo.jpg

Lab 4: Monitoring & Optimization

bash
# View logs
gcloud logging read "resource.type=gce_instance" --limit=50

# Query costs
bq query --use_legacy_sql=false 'SELECT service.description, ROUND(SUM(cost), 2) as cost FROM `PROJECT.billing_export.gcp_billing_export_v1` WHERE EXTRACT(MONTH FROM usage_start_time) = EXTRACT(MONTH FROM CURRENT_DATE()) GROUP BY service.description ORDER BY cost DESC'

# Auto-scaling metrics
gcloud compute instance-groups managed describe web-app-ig --zone=us-central1-a

Real-world Project Template

E-commerce Platform

Interview Questions

Practical Scenarios

Walk me through deploying a production web app from scratch.▾

1. Create instance template with startup script. 2. Create managed instance group (3 zones, 3 instances). 3. Create health checks (port 8080, /health). 4. Set auto-scaling (min 3, max 10, target 70% CPU). 5. Create backend service + URL map. 6. Create HTTP proxy + forwarding rule. 7. Get external IP. 8. Test by hitting IP in browser. 9. Generate load to watch scaling. Time: 15-20 minutes for 3-zone HA app.

How do you debug a slow database query?▾

1. Check Cloud SQL CPU/memory in Console. 2. Enable Query Insights (top 100 slowest queries). 3. Add indexes on frequently queried columns. 4. Use Cloud SQL Auth Proxy (encrypted connection). 5. Consider read replicas for scaling reads. 6. Cache results (Memorystore Redis). Expected improvement: 5-10x faster.

Your bill is $5k; reduce to $3k. How?▾

1. Right-size instances (use Recommender). 2. Buy commitments (25-35% off). 3. Use preemptible (70% off). 4. Aggressive auto-scaling. 5. Scale down non-prod at night (50% off). 6. Archive old data (95% off). 7. Cloud CDN for static assets. Achievable: 40% cost reduction in 2-3 months.

How do you ensure secure file uploads?▾

1. Use signed URLs (time-limited access). 2. Disable public bucket read. 3. Encrypt at rest (default on GCP). 4. Use Cloud DLP to scan sensitive data. 5. Enable Cloud Audit Logs for all access. 6. CORS only for approved origins. Result: Secure, auditable file upload flow.

Common Deployment Patterns

Pattern 1: Canary Deployment

Roll out new version to 5% of users first. Monitor error rates. If good, increase gradually. If bad, rollback instantly. Zero-downtime, safe deployments.

Pattern 2: Blue-Green Deployment

Run two identical environments (Blue, Green). Deploy new version to Green. Test thoroughly. Switch load balancer to Green instantly. If issues, switch back to Blue (instant rollback).

Pattern 3: Cost Optimization

Combine: right-sizing + commitments + preemptible for non-critical + scheduled scaling + storage archiving. Typical result: 30-50% cost savings.

Summary

Real-world deployment requires:

← Back to Course