DeepAudit/backend/docs/LOAD_TESTING.md

Load Testing Guide

Overview

This guide describes how to perform load testing on XCodeReviewer to ensure it can handle 100+ concurrent users.

Tools

• Locust: Python-based load testing tool
• Apache JMeter: Alternative load testing tool
• k6: Modern load testing tool

Installation

pip install locust

Load Test Scenarios

Scenario 1: API Endpoint Load Test

Create tests/load/test_api_load.py:

from locust import HttpUser, task, between
import random

class APILoadTest(HttpUser):
    wait_time = between(1, 3)
    
    def on_start(self):
        # Login
        response = self.client.post("/api/v1/auth/login", json={
            "username": "testuser",
            "password": "TestPass123!"
        })
        self.token = response.json()["access_token"]
        self.headers = {"Authorization": f"Bearer {self.token}"}
    
    @task(3)
    def list_projects(self):
        self.client.get("/api/v1/projects", headers=self.headers)
    
    @task(2)
    def list_tasks(self):
        self.client.get("/api/v1/tasks", headers=self.headers)
    
    @task(1)
    def get_statistics(self):
        self.client.get("/api/v1/statistics/overview", headers=self.headers)

Running Load Tests

Basic Load Test

locust -f tests/load/test_api_load.py --host=http://localhost:8000

Then open http://localhost:8089 and configure:

• Number of users: 100
• Spawn rate: 10 users/second

Headless Mode

locust -f tests/load/test_api_load.py \
    --host=http://localhost:8000 \
    --users 100 \
    --spawn-rate 10 \
    --run-time 5m \
    --headless

Performance Targets

• Response Time: 95th percentile < 500ms
• Throughput: > 1000 requests/second
• Error Rate: < 1%
• Concurrent Users: 100+

Monitoring During Load Tests

# Monitor system resources
htop

# Monitor database connections
psql -c "SELECT count(*) FROM pg_stat_activity;"

# Monitor Redis
redis-cli info stats

# Monitor API metrics
curl http://localhost:9090/metrics

Results Analysis

Expected results for 100 concurrent users:

• Average response time: < 200ms
• 95th percentile: < 500ms
• 99th percentile: < 1000ms
• Requests per second: > 1000
• Failure rate: < 1%

Task 28.2 Completion Checklist

• Install load testing tools
• Create load test scenarios
• Run tests with 100 concurrent users
• Monitor system resources
• Analyze results
• Document findings
• Optimize bottlenecks if needed
• Re-test after optimizations

Status: READY FOR EXECUTION

This task requires running the load tests in a staging environment.