feat(phase5-batch5.2): Load test suite with k6

[arpayidcloud-byte]

Summary

Implement Phase 5.2 - Load Tests per MASTER_PLAN_PRODUCTION.md

Changes

• Comprehensive k6 load test suite
• Performance test (100-1000 concurrent users)
• Graceful shutdown under load test
• Performance budgets defined
• Load test documentation
• npm scripts for running tests

Load Test Scenarios

Performance Test

7 Stages:

1. Ramp-up: 0 → 100 users (30s)
2. Steady: 100 users (1m)
3. Ramp-up: 100 → 500 users (30s)
4. Steady: 500 users (2m)
5. Ramp-up: 500 → 1000 users (30s)
6. Peak: 1000 users (3m)
7. Ramp-down: 1000 → 0 (30s)

User Behavior:

• 70% submit tasks
• 20% check status
• 10% list tasks

Graceful Shutdown Test

• 50 concurrent users
• Trigger SIGTERM mid-test
• Verify in-flight completion
• No data corruption

Performance Budgets

Metric	Threshold	Description
p95 Latency	< 500ms	95% of requests
p99 Latency	< 1000ms	99% of requests
Error Rate	< 1%	Failed requests

Usage

# Run performance test
pnpm test:load

# Run shutdown test
pnpm test:load:shutdown

# With custom URL
BASE_URL=http://localhost:3000 pnpm test:load

Files Added

• load-tests/performance.test.js (NEW - 250 lines)
• load-tests/shutdown.test.js (NEW - 120 lines)
• load-tests/README.md (NEW - 230 lines)
• package.json (MODIFIED - add scripts)
• README.md (MODIFIED - add load test commands)

CI Status

• Typecheck ✅
• Lint ✅
• Build ✅
• Load Tests ⏳ (requires running system)
• Documentation ✅

Related

• Phase 5.1: PR feat(phase5-batch5.1): Comprehensive E2E test suite #64 (E2E Tests) ✅ Merged
• Phase 5.2: This PR (Load Tests) ⬅️ CURRENT
• Phase 5.3: Security Tests (Next)

Next Steps

After merge:

1. Run load tests against staging environment
2. Tune performance based on results
3. Phase 5.3: Security Tests (SAST/DAST)

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.