| name | test-stress |
| description | Load, stress, spike, and soak testing — find where a service bends or breaks under pressure and where latency/error rates cross the line. Uses k6 (scriptable, low-overhead). Triggers on "load test", "stress test", "how much traffic can it handle", "test under pressure", "spike/soak test", "will it hold up at launch". Not for single-user functional checks (use test-blackbox/test-user-end). |
| allowed-tools | Read, Grep, Glob, Bash, Write, Edit |
test-stress — how it behaves under pressure
Push a real endpoint until you learn its limits: throughput, latency percentiles, and error rate as
concurrency climbs. Part of the wf-tester team. Uses k6 (a small, scriptable load tool); the QA
engine dispatches it as a CLI lane so results fold into the same CTRF/dashboard.
The four shapes
- Load — expected peak concurrency, held steady: does p95 latency and error rate stay within budget?
- Stress — ramp past expected peak until it degrades: find the breaking point and how it fails (graceful vs falls over).
- Spike — a sudden surge then drop: does it absorb and recover?
- Soak — moderate load for a long time: memory leaks, connection exhaustion, slow drift.
How
- Install k6 (
k6 version; on the CLI lane the engine calls it — set the target via MC_QA_TARGET_URL).
- Write a k6 script for the key journey (login → primary action), with thresholds that make it
pass/fail: e.g.
http_req_duration: ['p(95)<500'], http_req_failed: ['rate<0.01'].
- Start small and ramp:
k6 run --vus 10 --duration 30s script.js, then stages up to your target.
- Record throughput, p50/p95/p99, error rate, and the concurrency at first degradation. Feed the
summary into
.skull/qa/ so skull-dashboard charts it.
Safety
Load-test only systems you own or are authorized to test, and only staging/pre-prod unless
you've cleared production. High load can look like an attack and can cost real money (autoscaling) —
confirm the target and the ceiling with the owner first. Never point a load test at a third party.
