@kaizenreport/kensho-k6

Kensho summary helper for k6. k6 has its own JS runtime (Goja) so we can't ship a "reporter plugin" in the conventional sense — instead, you import this tiny self-contained module from your k6 script's handleSummary(data) and return the result. k6 writes the files for you.

Install

pnpm add -D @kaizenreport/kensho-k6 @kaizenreport/kensho

(Or, if you prefer not to bundle: copy dist/index.mjs next to your script and import it directly. The file is self-contained — zero external deps.)

Usage

// script.js
import http from 'k6/http';
import { check, sleep, group } from 'k6';
import { kenshoSummary } from '@kaizenreport/kensho-k6';

export const options = {
  scenarios: {
    smoke: {
      executor: 'constant-vus', vus: 5, duration: '10s',
      exec: 'smoke',
    },
    soak: {
      executor: 'ramping-vus', stages: [
        { duration: '5s', target: 10 },
        { duration: '20s', target: 10 },
      ],
      exec: 'soak',
    },
  },
  thresholds: {
    http_req_duration: ['p(95)<500'],
    http_req_failed:   ['rate<0.01'],
    checks:            ['rate>0.95'],
  },
};

export function smoke() {
  group('smoke / GET /products', () => {
    const r = http.get('https://test.k6.io/products');
    check(r, {
      'status is 200': (r) => r.status === 200,
      'body has items': (r) => r.body.includes('item'),
    });
  });
}

export function soak() {
  group('soak / GET /', () => {
    const r = http.get('https://test.k6.io/');
    check(r, { 'status is 200': (r) => r.status === 200 });
  });
}

export function handleSummary(data) {
  return kenshoSummary(data, {
    project: { name: 'API perf', slug: 'api-perf' },
  });
}

k6 run script.js
npx kensho generate
npx kensho open

What we capture

| k6 concept | Kensho mapping | | ----------------------------------------- | ------------------------------------------------- | | Each scenario in options.scenarios | One Kensho case named after the scenario | | Each check(...) inside a scenario | A sub-step on the scenario case (step.assertion populated on failure) | | Each threshold (e.g. http_req_duration: ['p(95)<500']) | One top-level case — pass if result.ok, fail otherwise. Labeled behavior.feature = 'thresholds'. | | Run-level metrics (http_req_duration, iterations, vus_max, data_sent, data_received, checks) | Flattened into run.env.vars (avg / p95 / count / rate keys) | | Per-iteration HTTP samples (opt-in via data.kenshoSamples) | step.request{} + step.response{} on the scenario, capped at opts.maxSteps (default 50) |

A scenario case is pass when all of its checks pass and fail when any check fails. A threshold case is pass when k6 reports result.ok === true and fail otherwise — these are surfaced as separate cases so the viewer's pass/fail counts include performance gates as first-class results.

Options

kenshoSummary(data, {
  project?: { name, slug, url },
  runId?: string,
  maxSteps?: number,             // default 50 — cap per-iteration samples
  output?: string,               // default 'kensho-results' (file-key prefix)
  framework?: { name, version },
  env?: object,                  // merged into run.env
})

Schema reference

The full Kensho v1 schema lives at @kaizenreport/kensho-schema. The summary object this helper returns matches the format the Kensho CLI's validate and generate commands expect.