Exploratory Load-Testing with AI Tooling - it's Time to Ramp Up!
summary
Demonstration shows AI-assisted exploratory load testing that actually runs tests and inspects artifacts: Test Cube plus an MCP server driving K6 from a cursor IDE, distributed runs on kind and minikube, aggregated TPS of ~7,000 and an attempted ramp to 10,000 that fails, and an injected sleep that drops performance by 99.89%. The system parses logs, suggests worker and config changes, can inspect source and propose fixes or pull requests, and succeeds sometimes but is non‑deterministic.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:30Exploratory load testing
- 5:47AI stack for exploratory load testing
- 8:14Demo setup with Test Cube and K6
- 12:11K6 calculator test demo
- 15:22Scaling to 10,000 TPS
- 19:21Cross-environment benchmarking
- 24:22Performance differences between environments
- 24:43Performance regression demo
- 26:30Detecting and diagnosing regression
- 29:11Challenges and limitations
- 30:56Benefits and future of AI-assisted testing
- 31:38Q&A
related talks
23mAI-Driven Load Test Analysis: From Routine Toward Action
42mEnhance Software Performance Testing Using Artificial Intelligence
01h 04mImpact of AI on #SoftwareTesting: Are Testers Ready? 🔮
43mExploratory Testing with AI
07mSmart Load Testing
55mAI-First Testing
48mSupercharged Testing: AI-Powered Workflows with Playwright + MCP
51mAI-Driven Strategies for Scalable & Resilient Performance Engg.