Performance Engineering

---

Performance Metrics

Why Percentiles Matter

---

Performance Laws

Little's Law

Amdahl's Law

Universal Scalability Law

---

Profiling

Types of Profilers

Flame Graphs

---

Load Testing

Load Testing Types

Load Test Example (k6)

// load-test.js
import http from 'k6/http';
import { check, sleep } from 'k6';

export const options = {
  stages: [
    { duration: '2m', target: 100 },   // Ramp up
    { duration: '5m', target: 100 },   // Steady state
    { duration: '2m', target: 200 },   // Stress
    { duration: '5m', target: 200 },   // Steady stress
    { duration: '2m', target: 0 },     // Ramp down
  ],
  thresholds: {
    http_req_duration: ['p(95)<200'],   // 95% < 200ms
    http_req_failed: ['rate<0.01'],     // Error rate < 1%
  },
};

export default function () {
  const res = http.get('https://api.example.com/orders');

  check(res, {
    'status is 200': (r) => r.status === 200,
    'response time < 200ms': (r) => r.timings.duration < 200,
  });

  sleep(1);
}

// Run: k6 run load-test.js

---

Optimization Techniques

Common Bottlenecks

Optimization Strategies

// 1. Caching
@Cacheable(value = "products", key = "#id")
public Product getProduct(Long id) {
    return productRepository.findById(id);
}

// 2. Connection Pooling
HikariConfig config = new HikariConfig();
config.setMaximumPoolSize(20);
config.setMinimumIdle(5);
config.setConnectionTimeout(30000);

// 3. Async Processing
@Async
public CompletableFuture<Report> generateReport(Long id) {
    return CompletableFuture.completedFuture(reportService.generate(id));
}

// 4. Batching
@Modifying
@Query("UPDATE Order o SET o.status = :status WHERE o.id IN :ids")
int updateStatusBatch(@Param("ids") List<Long> ids, @Param("status") Status status);

// 5. Pagination
public Page<Order> getOrders(int page, int size) {
    return orderRepository.findAll(PageRequest.of(page, size));
}

// 6. Index Optimization
@Table(indexes = {
    @Index(name = "idx_order_customer", columnList = "customer_id"),
    @Index(name = "idx_order_status_date", columnList = "status, created_at")
})
public class Order { }

---

JVM Performance

GC Tuning

---

Capacity Planning

---

Common Interview Questions

1. Why use percentiles instead of average?
2. Average hides outliers
3. P99 shows worst-case user experience

4. Tail latency matters for user satisfaction

5. What is Amdahl's Law?

6. Limits speedup from parallelization
7. Serial portion limits max improvement

8. 10% serial = max 10x speedup

9. How to identify performance bottlenecks?

10. Profile CPU, memory, I/O
11. Check resource utilization
12. Analyze flame graphs

13. Use APM tools

14. Load test types?

15. Smoke: Verify works
16. Load: Expected traffic
17. Stress: Find breaking point

18. Soak: Long duration

19. How to improve latency?

20. Caching
21. Connection pooling
22. Async processing
23. Database optimization
24. CDN for static content

---

• *