🚀 Qnova.uk Performance Analysis Report

Project: Load Testing & Scalability Audit
Author: Muhammad Nasir Dantiye Tools: Apache JMeter 5.6.3 (CLI Mode), Git, macOS

📊 Executive Summary

This project evaluates the performance and stability of the qnova.uk web server. While the server showed excellent stability up to 500 users, the local testing environment reached a physical thread limit at 2,016 concurrent users.

🛠 Test Configuration

Parameter	Value
Target URL	https://qnova.uk
Protocol	HTTPS / GET
Ramp-up Period	60 - 300 Seconds
Environment	macOS (Local CLI Mode)

📈 Test Results

Test Phase	Threads (Users)	Avg Response Time	Error %	Status
Baseline	50	1253ms	0.00%	✅ PASSED
Scalability	100	1258ms	0.00%	✅ PASSED
Stress Test	500	1331ms	0.00%	✅ PASSED
Peak Load	3,000	916ms*	73.83%*	⚠️ LIMIT REACHED

🔍 Bottleneck Analysis

• Hardware Limitation: During the 3,000-user peak test, the local machine failed to spawn new threads after reaching 2,016 active users.
• Error Log: pthread_create failed (EAGAIN) - This indicates the macOS kernel limit for processes/threads was reached.
• Server Response: Up until the hardware crash, the server (qnova.uk) maintained an average response time of under 1.4s, suggesting high server-side efficiency.

📝 Recommendations

1. Distributed Testing: To test the full 3,000-user capacity, a "Master-Slave" JMeter configuration using multiple load generators is required.
2. Kernel Tuning: For further local testing, macOS maxfiles and maxproc limits should be increased via sysctl.