writing-k6-tests

Write complete, runnable k6 scripts for load testing or Grafana Synthetic Monitoring (SM) checks. You have no k6 execution tools — authoring is the deliverable; give the user the exact run command.

Decide the goal first

• Load/perf test (throughput, latency under load, breaking point): full k6 script with load profile. This is the default reading of "load test".
• Synthetic Monitoring check (uptime, "does the user journey work right now"): one iteration, one VU, fixed schedule. No load idioms — vus, stages, iterations are ignored by SM and thresholds are not supported. Say "check", "probe", "execution" — never "VUs" or "ramp".

A script can serve both, but the options shape differs. Ask only if the goal is genuinely unclear.

Load-test skeleton (start here, adapt endpoints and profile)

import http from 'k6/http';
import { check, sleep } from 'k6';
import { Trend, Rate } from 'k6/metrics';

const latency = new Trend('checkout_latency');
const errors = new Rate('checkout_errors');

export const options = {
  scenarios: {
    load: {
      executor: 'ramping-arrival-rate', // open model: pace stays honest as latency degrades
      startRate: 10, timeUnit: '1s',
      preAllocatedVUs: 50, maxVUs: 200,
      stages: [
        { target: 50, duration: '2m' },
        { target: 50, duration: '5m' },
        { target: 0, duration: '1m' },
      ],
    },
  },
  thresholds: {
    http_req_duration: ['p(95)<500'],
    http_req_failed: ['rate<0.01'],
  },
};

export default function () {
  const res = http.get('https://example.com/api/checkout');
  check(res, { 'status 200': (r) => r.status === 200 });
  latency.add(res.timings.duration);
  errors.add(res.status >= 500);
  sleep(1); // user think time — required in VU-based (closed-model) tests
}

Executor cheat sheet