api-load-testing

# API Load Testing & Concurrency Benchmark (Ibra X 3.1) Alat pengujian beban endpoint REST/GraphQL berkecepatan tinggi tanpa dependensi eksternal yang rumit, dibangun menggunakan Python `asyncio` dan `urllib`. ## 1. Quick Async Load Tester (Python) Simpan skrip berikut sebagai `load_test.py`: ```python #!/usr/bin/env python3 import asyncio, time, urllib.request, json, statistics, sys async def worker(url, headers, duration_sec, results): end_time = time.time() + duration_sec while time.time() < end_time: t0 = time.time() try: req = urllib.request.Request(url, headers=headers) with urllib.request.urlopen(req, timeout=10) as resp: status = resp.status resp.read() elapsed = time.time() - t0 results.append((elapsed, status, None)) except Exception as e: elapsed = time.time() - t0 results.append((elapsed, 0, str(e))) await asyncio.sleep(0.001) async def run_benchmark(url, concurrency=20, duration=10, headers=None): if headers is None: headers = {"User-Agent": "Ibra-LoadTester/3.1"} print(f"🚀 Memulai benchmark target: {url}") print(f" Konkurensi: {concurrency} workers | Durasi: {duration} detik") results = [] tasks = [worker(url, headers, duration, results) for _ in range(concurrency)] start = time.time() await asyncio.gather(*tasks) total_time = time.time() - start latencies = [r[0] * 1000 for r in results if r[1] == 200] errors = [r for r in results if r[1] != 200] total_reqs = len(results) rps = total_reqs / total_time if total_time > 0 else 0 print("\n" + "=" * 50) print("📊 HASIL PENGUJIAN BEBAN:") print("=" * 50) print(f"Total Request : {total_reqs}") print(f"Request Berhasil (200): {len(latencies)}") print(f"Request Gagal/Error : {len(errors)}") print(f"Throughput Rata-rata : {rps:.1f} Req/Sec") if latencies: latencies.sort() p50 = statistics.median(latencies) p90 = latencies[int(len(latencies) * 0.90)] p95 = latencies[int(len(latencies) * 0.95)] p99 = latencies[int(len(latencies) * 0.99)] print(f"Latensi Rata-rata : {statistics.mean(latencies):.2f} ms") print(f"Latensi Min / Max : {min(latencies):.2f} ms / {max(latencies):.2f} ms") print(f"Latensi P50 (Median) : {p50:.2f} ms") print(f"Latensi P90 : {p90:.2f} ms") print(f"Latensi P95 : {p95:.2f} ms") print(f"Latensi P99 : {p99:.2f} ms") print("=" * 50) if __name__ == "__main__": target = sys.argv[1] if len(sys.argv) > 1 else "https://httpbin.org/get" conc = int(sys.argv[2]) if len(sys.argv) > 2 else 10 dur = int(sys.argv[3]) if len(sys.argv) > 3 else 5 asyncio.run(run_benchmark(target, conc, dur)) ``` ## 2. Cara Menjalankan ```bash # Uji target dengan 20 koneksi konkuren selama 10 detik python3 load_test.py https://api.anda.com/v1/health 20 10 ``` ## 3. Metrik Utama untuk Evaluasi - **P95 Latency**: 95% pengguna merasakan latensi di bawah nilai ini. Harus berada di bawah 200ms untuk API real-time. - **P99 Latency**: Indikator penumpukan antrean koneksi (*cold-start* atau *tail latency*). - **Error Rate**: Harus 0.00% pada beban puncak. Jika muncul status `429` (Too Many Requests) atau `502/504` (Bad Gateway), berarti server membutuhkan auto-scaling atau reverse proxy tuning.