Automated smoke load tests in CI

Severity: Optional

Quick, lightweight load tests running in CI catch obvious regressions automatically. Not full production-scale tests, but sanity checks that new code doesn't completely break under any load.

Check automatically:

1. Check CI configs for load test jobs:

# Check GitHub Actions
grep -riE "k6|artillery|loadtest|autocannon|smoke.*test|performance.*test" .github/workflows/ --include="*.yml" --include="*.yaml" 2>/dev/null

# Check other CI systems
grep -riE "k6|artillery|loadtest|performance" .circleci/ .gitlab-ci.yml Jenkinsfile bitbucket-pipelines.yml 2>/dev/null

# Look for load test npm scripts
grep -E "\"(load|stress|smoke|perf).*test" package.json 2>/dev/null

# Check for k6 cloud or Artillery Pro CI integrations
grep -riE "K6_CLOUD|ARTILLERY_CLOUD|k6.*run.*--cloud" .github/workflows/ --include="*.yml" 2>/dev/null

# Look for performance test in CI job names
grep -riE "job:.*perf|job:.*load|name:.*performance|name:.*load.*test" .github/workflows/ --include="*.yml" 2>/dev/null

Ask user:

• "Do you run any load tests automatically in CI?"
• "What level of load? (smoke = light sanity check vs. full = production-like)"
• "Do CI load tests block deploys on failure?"

Smoke test characteristics:

• Low concurrency (10-50 virtual users)
• Short duration (30-60 seconds)
• Key endpoints only (health, main API routes)
• Fast feedback (< 5 minutes)

Cross-reference with:

• LST-001 (tool needed)
• LST-003 (full load tests complement CI smoke tests)
• Section 5 (CI/CD pipeline) - load tests as pipeline stage

Pass criteria:

• Smoke-level load tests run in CI (quick, low concurrency)
• Tests catch obvious regressions (10x slower response, errors under light load)
• Results visible in CI output or dashboard

Fail criteria:

• No automated load tests (all manual)
• Tests exist but don't run in CI
• Tests run but nobody looks at results

Notes: This is Optional because not all projects need CI-integrated load tests. For high-traffic production systems, it's valuable. For internal tools or early-stage products, manual load testing (LST-003) may be sufficient.

Evidence to capture:

• CI job name/location running load tests
• Concurrency level and duration
• Whether failures block deployment

---