website-enchun-mgr/apps/backend/tests/k6/TESTING-GUIDE.md

Load Testing Execution Guide

Test Execution Checklist

Pre-Test Requirements

• Backend server is running (pnpm dev)
• Database is accessible
• k6 is installed (k6 version)
• Environment variables are configured

Verification Test

Run the verification script first to ensure everything is set up correctly:

k6 run tests/k6/verify-setup.js

Expected Output:

=== K6 Setup Verification ===
Target: http://localhost:3000
Home page status: 200
API status: 200
Pages API response time: 123ms
✓ Server is reachable
✓ Home page responds
✓ API endpoint responds

Test Scenarios

1. Public Browsing Test

Purpose: Verify public pages can handle 100 concurrent users

Prerequisites:

• Backend running
• Public pages accessible

Execution:

# Local testing
pnpm test:load

# With custom URL
k6 run --env BASE_URL=http://localhost:3000 tests/k6/public-browsing.js

# Staging
k6 run --env BASE_URL=https://staging.enchun.tw tests/k6/public-browsing.js

Success Criteria:

• p95 response time < 500ms
• Error rate < 1%
• 100 concurrent users sustained for 2 minutes

What It Tests:

• Homepage rendering
• Page navigation
• Static content delivery
• Database read operations

---

2. API Performance Test

Purpose: Verify API endpoints meet performance targets

Prerequisites:

• Backend running
• API endpoints accessible

Execution:

# Local testing
pnpm test:load:api

# With custom URL
k6 run --env BASE_URL=http://localhost:3000 tests/k6/api-performance.js

# Staging
k6 run --env BASE_URL=https://staging.enchun.tw tests/k6/api-performance.js

Success Criteria:

• p95 response time < 300ms
• Error rate < 0.5%
• Throughput > 100 req/s

What It Tests:

• REST API endpoints
• GraphQL queries
• Authentication endpoints
• Concurrent API requests

---

3. Admin Operations Test

Purpose: Verify admin panel can handle 20 concurrent users

Prerequisites:

• Backend running
• Valid admin credentials

Execution:

# Local testing
k6 run \
  --env ADMIN_EMAIL=admin@enchun.tw \
  --env ADMIN_PASSWORD=yourpassword \
  tests/k6/admin-operations.js

# Or use npm script
ADMIN_EMAIL=admin@enchun.tw ADMIN_PASSWORD=yourpassword \
  pnpm test:load:admin

Success Criteria:

• p95 response time < 700ms
• Error rate < 1%
• 20 concurrent users sustained for 3 minutes

What It Tests:

• Login/authentication
• CRUD operations
• Admin panel performance
• Database write operations

Warning: This test creates draft posts in the database. Clean up manually after testing.

---

Test Execution Strategy

Phase 1: Development Testing

Run tests locally during development with low load:

# Quick smoke test (10 users)
k6 run --env STAGED_USERS=10 tests/k6/public-browsing.js

# API smoke test (5 users)
k6 run --env STAGED_USERS=5 tests/k6/api-performance.js

Phase 2: Pre-Deployment Testing

Run full test suite against staging:

# Run all tests
pnpm test:load:all

# Or individual tests with full load
k6 run --env BASE_URL=https://staging.enchun.tw tests/k6/public-browsing.js
k6 run --env BASE_URL=https://staging.enchun.tw tests/k6/api-performance.js
k6 run --env BASE_URL=https://staging.enchun.tw \
  --env ADMIN_EMAIL=$STAGING_ADMIN_EMAIL \
  --env ADMIN_PASSWORD=$STAGING_ADMIN_PASSWORD \
  tests/k6/admin-operations.js

Phase 3: Production Monitoring

Schedule automated tests (via GitHub Actions or cron):

• Daily: Public browsing and API tests
• Weekly: Full test suite including admin operations
• On-demand: Before major releases

---

Result Analysis

Key Metrics to Check

1. p95 Response Time

   • Public pages: < 500ms ✅
   • API endpoints: < 300ms ✅
   • Admin operations: < 700ms ✅

2. Error Rate

   • Public: < 1% ✅
   • API: < 0.5% ✅
   • Admin: < 1% ✅

3. Throughput

   • API: > 100 req/s ✅
   • Pages: > 50 req/s ✅

4. Virtual Users (VUs)

   • Should sustain target VUs for full duration
   • No drops or connection errors

Sample Output Analysis

✓ http_req_duration..............: avg=185ms p(95)=420ms
✓ http_req_failed................: 0.00% ✓ 0      ✗ 12000
✓ checks.........................: 100.0% ✓ 12000 ✗ 0
iterations.....................: 2400   20  /s
vus............................: 100    min=100    max=100

This result shows:

• ✅ p95 = 420ms (< 500ms threshold)
• ✅ Error rate = 0% (< 1% threshold)
• ✅ All checks passed
• ✅ Sustained 100 VUs

---

Troubleshooting Guide

Issue: Connection Refused

Symptoms:

ERRO[0000] GoError: dial tcp 127.0.0.1:3000: connect: connection refused

Solutions:

1. Ensure backend is running: pnpm dev
2. Check port is correct: --env BASE_URL=http://localhost:3000
3. Verify firewall isn't blocking connections

---

Issue: High Error Rate (> 1%)

Symptoms:

✗ http_req_failed................: 2.50% ✓ 11700  ✗ 300

Common Causes:

1. Server overloaded → Reduce VUs
2. Database connection issues → Check DB logs
3. Missing authentication → Check credentials
4. Invalid URLs → Verify BASE_URL

Solutions:

# Reduce load to diagnose
k6 run --env STAGED_USERS=10 tests/k6/public-browsing.js

# Check server logs in parallel
tail -f logs/backend.log

---

Issue: Slow Response Times (p95 >= 500ms)

Symptoms:

http_req_duration..............: avg=450ms p(95)=850ms

Common Causes:

1. Unoptimized database queries
2. Missing database indexes
3. Large payload sizes
4. Server resource constraints

Solutions:

1. Check database query performance
2. Review database indexes
3. Optimize images/assets
4. Scale server resources
5. Enable caching

---

Issue: Login Failed (Admin Test)

Symptoms:

✗ login successful
status: 401

Solutions:

1. Verify admin credentials are correct
2. Check admin user exists in database
3. Ensure admin user has correct role
4. Try logging in via admin panel first

---

Performance Optimization Tips

Based on test results, you may need to:

1. Add Database Indexes

   // Example: Index on frequently queried fields
   Posts.createIndex({ title: 1 });
   Posts.createIndex({ status: 1, createdAt: -1 });
   

2. Enable Caching

   • Cache global API responses
   • Cache public pages
   • Use CDN for static assets

3. Optimize Images

   • Use WebP format
   • Implement lazy loading
   • Serve responsive images

4. Database Optimization

   • Limit query depth where possible
   • Use projection to reduce payload size
   • Implement pagination

5. Scale Resources

   • Increase server memory/CPU
   • Use database connection pooling
   • Implement load balancing

---

Reporting

Generate Reports

# JSON report
k6 run --out json=results.json tests/k6/public-browsing.js

# HTML report
npm install -g k6-reporter
k6-reporter results.json --output results.html
open results.html

Share Results

Include in your PR/daily standup:

• p95 response time
• Error rate
• Throughput
• Any issues found

---

Best Practices

1. Run tests regularly - Catch regressions early
2. Test on staging first - Avoid breaking production
3. Monitor trends - Track performance over time
4. Share results - Keep team informed
5. Update baselines - Adjust as system evolves
6. Don't ignore failures - Investigate all issues

---

Last Updated: 2025-01-31 Owner: QA Team