Skip to content

k6-documentation.md

Latest commit

 

History

History
241 lines (181 loc) · 8.61 KB

k6-documentation.md

File metadata and controls

241 lines (181 loc) · 8.61 KB
Raw

k6 Load and Stress Testing

This document provides comprehensive information about the k6 load and stress testing setup for the Frequency Developer Gateway microservices.

Overview

The Gateway project includes extensive k6 load and stress testing capabilities across all microservices. These tests help ensure the reliability, performance, and scalability of the Gateway services under various load conditions.

Prerequisites

Before running k6 tests, ensure you have:

  1. k6 installed: Download from k6.io
  2. Gateway services running: Use ./start.sh to start all services locally
  3. Node.js (version 18 or higher): Required for generating test data and running helper scripts

Port Configuration

The Gateway services use different ports when running in Docker containers:

Service Internal Port External Port (Docker) Purpose
Content Publishing API 3000 3010 Main API endpoints
Content Publishing Worker 3000 3020 Background processing
Content Watcher 3000 3011 Content monitoring
Graph API 3000 3012 Social graph operations
Graph Worker 3000 3022 Background processing
Account API 3000 3013 Account management

Note: The k6 tests are configured to target the appropriate ports:

  • Standard tests (health checks, v1 endpoints) target port 3000
  • v2 batch announcement tests target port 3010 (Docker external port)
  • File upload tests target port 3000 (internal container port)

Available Test Suites

Account Service (apps/account-api/k6-test/)

Test File Description Purpose Port
health-check.k6.js Basic health check endpoint testing Verify service availability 3000
new-sign-up.k6.js User registration flow testing Test account creation performance 3000
account-service-load.k6.js Comprehensive API load testing Full service load testing (currently skipped) 3000
signups.gen.js Generated test data Pre generated signup payloads N/A

Key Features:

  • Test data generation with npm run generate:signup
  • Realistic signup payloads to avoid blockchain interaction
  • Health check validation before main tests

Content Publishing Service (apps/content-publishing-api/k6-test/)

Test File Description Purpose Port
health-check.k6.js Basic health check endpoint testing Verify service availability 3000
script.k6.js Standard API load testing General service performance 3000
batch-announcement-load.k6.js Batch content publishing load testing Test batch announcement performance 3010
batch-announcement-stress.k6.js Advanced stress testing Comprehensive stress scenarios 3010
script_sm_files.k6.js Small file upload testing Test small file handling 3000
script_md_files.k6.js Medium file upload testing Test medium file handling 3000
script_lg_files.k6.js Large file upload testing Test large file handling 3000

Key Features:

  • Multiple load scenarios (light, medium, heavy, burst)
  • File size-specific testing
  • Realistic batch data generation
  • Comprehensive stress testing with multiple phases

Port Configuration:

  • Port 3000: Standard API endpoints (v1), health checks, and file upload tests
  • Port 3010: v2 batch announcement endpoints (mapped from Docker container port)

Content Watcher Service (apps/content-watcher/k6-test/)

Test File Description Purpose Port
health-check.k6.js Basic health check endpoint testing Verify service availability 3000

Key Features:

  • Simple health check validation
  • Service availability testing

Graph Service (apps/graph-api/k6-test/)

Test File Description Purpose Port
health-check.k6.js Basic health check endpoint testing Verify service availability 3000
script.js Standard API load testing General service performance 3000

Key Features:

  • Social graph operation testing
  • Relationship management performance testing

Test Scenarios

Load Testing Scenarios

The content publishing service includes predefined load testing scenarios:

Light Load

  • VUs: 5
  • Duration: 30 seconds
  • Purpose: Basic functionality testing
  • Thresholds: 99% success rate, less than 2 seconds average response time

Medium Load

  • VUs: 20
  • Duration: 60 seconds
  • Purpose: Normal operation testing
  • Thresholds: 98% success rate, less than 5 seconds average response time

Heavy Load

  • VUs: 200
  • Duration: 120 seconds
  • Purpose: Stress testing
  • Thresholds: 90% success rate, less than 10 seconds average response time

Burst Load

  • Stages: Ramp up → Spike → Ramp down
  • Purpose: Spike testing
  • Thresholds: 90% success rate, less than 15 seconds average response time

Stress Testing Phases

The stress testing includes multiple phases:

  1. Ramp-up Phase: Gradual increase from 1 to 10 VUs over 90 seconds
  2. Sustained Load: Constant 10 VUs for 2 minutes
  3. Spike Testing: Arrival rate testing with up to 25 requests per second
  4. Burst Testing: High-intensity burst of 50 requests per second for 1 minute

Running Tests

Basic Health Check

cd apps/[service-name]/k6-test
k6 run health-check.k6.js

Load Testing with Specific Scenario

cd apps/content-publishing-api/k6-test
SCENARIO=heavy k6 run batch-announcement-load.k6.js

Stress Testing

cd apps/content-publishing-api/k6-test
k6 run batch-announcement-stress.k6.js

File Size Testing

cd apps/content-publishing-api/k6-test
k6 run script_sm_files.k6.js  # Small files
k6 run script_md_files.k6.js  # Medium files
k6 run script_lg_files.k6.js  # Large files

Test Data Generation

Account Service

cd apps/account-api/k6-test
npm run generate:signup

This generates 100 valid signup payloads in signups.gen.js to avoid blockchain interaction during testing.

Content Publishing Service

The service includes helper functions in helpers.js for generating realistic test data:

  • createRealisticBatchData(): Creates realistic batch announcement data
  • createMultipartBatchData(): Creates multipart form data for file uploads
  • Various file size scenarios and error conditions

Configuration Options

Environment Variables

  • SCENARIO: Select load testing scenario (light, medium, heavy, burst)
  • BASE_URL: Override default service URL (defaults to localhost)

k6 Options

Each test file includes configurable options:

  • VUs
  • Test duration
  • Performance thresholds
  • Connection settings
  • Custom metrics

Performance Thresholds

Standard Thresholds

  • Success Rate: 90% to 99% depending on scenario
  • Response Time: Less than 2 to 15 seconds average depending on scenario
  • Error Rate: Less than 1% to 20% depending on scenario
  • Request Rate: Varies by test type

Custom Metrics

  • Successful requests counter
  • Failed requests counter
  • Response time trends
  • Scenario-specific error rates

CI/CD Integration

The project includes CI/CD tools in tools/ci-k6/ for automated testing:

  • Chain setup scripts for testnet testing
  • Automated test execution
  • Performance reporting

Best Practices

  1. Start with Health Checks: Always run health checks before load testing
  2. Use Realistic Data: Generate test data that mimics real usage patterns
  3. Monitor Resources: Watch system resources during stress testing
  4. Gradual Scaling: Start with light loads and gradually increase
  5. Environment Isolation: Run tests against dedicated test environments
  6. Result Analysis: Review k6 output for performance insights

Troubleshooting

Common Issues

  1. Connection Errors: Ensure services are running and accessible
  2. Port Configuration Errors: Verify you're targeting the correct port for each test type
    • Standard API tests: Port 3000
    • v2 batch announcement tests: Port 3010
    • File upload tests: Port 3000
  3. EOF Errors: May occur with file uploads; tests include connection settings to handle these
  4. Timeout Errors: Adjust k6 timeouts for large file uploads
  5. Memory Issues: Monitor system resources during heavy load testing

Debug Mode

Run tests with verbose output:

k6 run --verbose health-check.k6.js

For more information about k6, visit the official documentation.