Performance / Scenario
The Load Test That Lied
A k6 script reports 20ms p95 and 'all green', but production melts at launch. Read the script and find why the test was meaningless.
- Difficulty
- Hard
- Format
- Scenario
- Points
- 200
- Estimate
- 14 min
// MISSION BRIEF
Your Mission
A team load-tested their checkout with k6, saw a beautiful 20ms p95, and shipped. Black Friday, it fell over at 200 concurrent users. The load test lied.
Read the k6 script and the results, then diagnose why the test gave false confidence, and what a real test would do.
// FIRST CONTACT
Battle teaser
First artifact
load-test.js (k6)
Why did the load test give false confidence?
- A200 VUs is too few to matter
- BIt only exercised a CDN-cached static page, never the checkout path that actually has backend/DB cost
- C30 seconds is too long a duration
- Dk6 can't measure p95 accurately
Answers, scoring, hints, and the full battle stay sealed.
// SKILL TAGS
performancek6load-testingscenario