4.4 KiB

Raw Blame History

Load Testing Guide for C Programming Learning Management System

This guide explains how to run load tests on the LMS-C application using Locust with different scenarios and configurations.

Prerequisites

Before running load tests, ensure you have:

Podman installed on your system
The LMS-C application is properly configured with content and tokens
The tokens_siswa.csv file contains student tokens (the load testing script will use these for realistic user simulation)

Running Load Tests

1. Basic Load Test

To run a basic load test with the default configuration:

# Navigate to the test directory
cd /path/to/lms-c/elemes/test

# Run the load test with default settings (10 simulated users)
podman-compose -f podman-compose.yml up --build

Then access the Locust web interface at http://localhost:8089 and configure your test parameters.

2. Distributed Load Test

For larger-scale testing, you can run a distributed test with multiple workers:

# Start the master node
podman-compose -f podman-compose.yml up --scale worker=3 master

# Or run with specific environment variables
TARGET_URL=http://your-lms-url.com LOCUST_NUM_STUDENTS=100 podman-compose -f podman-compose.yml up --scale worker=3 master

This will start 1 master and 3 worker nodes to distribute the load.

3. Environment Variables

The load testing configuration can be customized using these environment variables:

TARGET_URL: The URL of the LMS-C application to test (default: http://example.com)
LOCUST_NUM_STUDENTS: Number of simulated students/users (default: 10, but will automatically detect from tokens_siswa.csv)

Example:

TARGET_URL=http://192.168.1.100:5000 LOCUST_NUM_STUDENTS=50 podman-compose -f podman-compose.yml up --build

4. Customizing Test Scenarios

The locustfile.py implements several user behavior patterns:

WebsiteUser Class

Simulates basic website visitors
Performs tasks like viewing homepage, lessons, and compiling code
Weight: 1 (less frequent)

AdvancedUser Class

Simulates engaged students who actively participate
Performs more complex behaviors like lesson navigation and intensive code compilation
Weight: 2 (twice as likely to be chosen as WebsiteUser)

Task Distribution

view_homepage: Weight 3 (most common action)
compile_code: Weight 4 (very common action)
view_lesson: Weight 2 (common action)
login_student: Weight 1 (less frequent but important)
validate_token: Weight 1 (essential for tracking)
track_progress: Weight 1 (important for completion tracking)

5. Realistic Student Simulation

The load testing script reads from tokens_siswa.csv to simulate real students:

Each simulated user gets assigned a real student token from the CSV
This ensures realistic progress tracking behavior
The number of simulated users should match or be proportional to the number of tokens in the CSV

6. Monitoring and Analysis

Access the Locust web interface at http://localhost:8089 to:

Configure the number of users and spawn rate
Monitor real-time statistics
View response times, failure rates, and throughput
Download test reports

7. Running Without the Web Interface

You can also run Locust in headless mode:

# Run with specific parameters without web UI
podman run -v $(pwd)/..:/mnt/locust \
  -e LOCUST_HOST=http://your-target-url.com \
  -e LOCUST_USERS=100 \
  -e LOCUST_SPAWN_RATE=10 \
  -e LOCUST_RUN_TIME=10m \
  locustio/locust -f /mnt/locust/elemes/test/locustfile.py --headless

8. Scaling Recommendations

For 1-50 concurrent users: Single master node is sufficient
For 50-200 concurrent users: Use 1 master + 2-3 worker nodes
For 200+ concurrent users: Scale workers proportionally (1 master + 5+ workers)

9. Best Practices

Always test against a staging environment that mirrors production
Gradually increase the number of users to identify performance bottlenecks
Monitor server resources (CPU, memory, disk I/O) during tests
Run tests multiple times to account for variations
Clean up resources after testing to avoid unnecessary resource consumption

10. Troubleshooting

If you encounter issues:

Ensure the target LMS-C application is accessible from the Locust containers
Check that the tokens_siswa.csv file is properly mounted and readable
Verify that the content directory has lesson files for realistic testing
Monitor container logs with podman logs -f <container-name>

4.4 KiB Raw Blame History