Constant RPS Load Test

This directory contains a Python script for conducting load tests with a fixed requests per second (RPS) rate.

Prerequisites

Before running the load test, ensure you have the following installed:

• Python 3.6+ : You need Python version 3.6 or higher
• Required Python packages:
  • requests
  • statistics
  • zoneinfo (for Python < 3.9, install backports.zoneinfo)

Installation

1. Install required packages:

   pip install requests statistics

2. For Python < 3.9, also install zoneinfo:

   pip install backports.zoneinfo

Preparing Test Payloads

1. Create a directory named payloads in the same folder as the script
2. Add JSON files in the payloads directory representing the request bodies to be sent
   • The script will randomly select from these payloads for each request

Running the Test

To start the load test, run the following command:

python3 load_test_rps.py <target_url> <duration_seconds> <requests_per_second>

Example:

python3 load_test_rps.py https://api.example.com/endpoint 600 160

This will run a test against the specified URL for 600 seconds (10 minutes) at 160 requests per second.

Command Line Arguments

The script requires three command line arguments:

• target_url: The endpoint URL to send requests to
• duration_sec: Duration of the test in seconds
• rps: Target requests per second to maintain

Monitoring the Test

During execution, the script displays progress updates every 500 requests showing:

• Current timestamp
• Number of requests sent (and total planned)
• Current average RPS
• Response status distribution (200, 500, other)
• Current latency metrics (TP90, TP95, TP99)

Test Summary

After completion, the script outputs a summary including:

• Total requests sent
• Response status counts
• Latency percentiles (TP90, TP95, TP99)
• Median response time
• Average RPS achieved
• Actual test duration

Key Features

• Maintains constant RPS regardless of response times
• Uses threading for concurrent requests
• Handles timing compensation to ensure steady request rate
• Collects and reports detailed latency metrics
• Randomly selects from multiple payload options
• Provides real-time progress updates