[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6KW5K2T6FF7MGJ5QJQP9PX3":3,"skill-same:01M6KW5K2T6FF7MGJ5QJQP9PX3":68},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":23,"publishedAt":42,"linkOnly":43,"markdownContent":44,"markdownBody":45,"frontMatter":46,"refName":47,"commitSha":48,"skillDirectory":49,"forksCount":14,"repositoryHtmlUrl":50,"repositoryHomepageUrl":6,"ownerHtmlUrl":51,"ownerBlogUrl":6,"ownerDescription":52,"updatedAt":53,"install":54,"resources":67},"01M6KW5K2T6FF7MGJ5QJQP9PX3","k6-load-testing",null,"Invoke this skill when the task is to validate that the system sustains its SLOs at **realistic production throughput over time** — including endurance (soak) testing for memory leaks, connection pool exhaustion, and GC pressure.","rafsanulhasan\u002FHypermediaEngine","rafsanulhasan","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F2341142?v=4","01M6KW2HWPTZWK051N7B0HTBVG","Rafsanul Hasan","01M6KW2J0N3613382MF5DRN79H",0,"GPL-3.0","https:\u002F\u002Fgithub.com\u002Frafsanulhasan\u002FHypermediaEngine\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fk6-load-testing\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41],"load testing","load","k6","k6 load testing","testing","gc","for","at","pool","connection","throughput","time","task","soak","system","Production","validate","memory","2026-11-10 10:12:30",false,"---\nname: k6-load-testing\ndescription: \"Invoke this skill when the task is to validate that the system sustains its SLOs at **realistic production throughput over time** — including endurance (soak) testing for memory leaks, connection pool exhaustion, and GC pressure.\"\n---\n\n# k6 Load Testing\n\n## When to Use\nInvoke this skill when the task is to validate that the system sustains its SLOs at **realistic production throughput over time** — including endurance (soak) testing for memory leaks, connection pool exhaustion, and GC pressure.\n\nThis skill covers:\n- **Average-load tests**: validate SLOs at expected normal traffic (use for regression gates)\n- **Soak \u002F endurance tests**: validate reliability under sustained load for hours\n- **Multi-scenario weighted load**: model realistic endpoint traffic mixes\n\nFor above-normal load, use `k6-stress-testing`. For basic smoke\u002Fbaseline validation, use `k6-performance-testing`.\n\n**Always run in this order**: smoke → average-load → stress → soak. Never run a soak test on a system that hasn't first passed average-load and stress.\n\n## Executor Selection\n\n| Executor | Model | When to Use |\n|----------|-------|-------------|\n| `constant-vus` | Closed | Specific concurrent user count; throughput varies with latency |\n| `constant-arrival-rate` | **Open** | Fixed RPS regardless of response time. **Preferred for REST API load tests** |\n| `per-vu-iterations` | Closed | Partition fixed test data across VUs (each VU processes its own slice) |\n\n**Prefer `constant-arrival-rate`** for REST API load tests. In closed models, slow responses reduce RPS invisibly — the system appears healthy while throughput degrades.\n\n## Constant Arrival Rate Load Test (REST API)\n\n```javascript\nimport http from 'k6\u002Fhttp';\nimport { check, sleep } from 'k6';\nimport { SharedArray } from 'k6\u002Fdata';\n\nconst users = new SharedArray('users', function () {\n  return JSON.parse(open('.\u002Ftest-data\u002Fusers.json'));\n});\n\nexport const options = {\n  scenarios: {\n    constant_load: {\n      executor: 'constant-arrival-rate',\n      rate: 50,                 \u002F\u002F 50 iterations per second\n      timeUnit: '1s',\n      duration: '10m',\n      preAllocatedVUs: 20,      \u002F\u002F initial VU pool\n      maxVUs: 100,              \u002F\u002F hard ceiling\n    },\n  },\n  thresholds: {\n    http_req_failed:   ['rate\u003C0.01'],          \u002F\u002F \u003C1% error rate\n    http_req_duration: [\n      'p(90)\u003C400',\n      'p(95)\u003C800',\n      'p(99)\u003C2000',\n    ],\n    http_req_waiting:  ['p(95)\u003C600'],          \u002F\u002F TTFB SLO\n    checks:            ['rate>0.99'],\n    dropped_iterations: ['count\u003C10'],          \u002F\u002F near-zero tolerance for dropped iterations\n  },\n};\n\nexport default function () {\n  const user = users[Math.floor(Math.random() * users.length)];\n  const res = http.get(`${__ENV.BASE_URL}\u002Fapi\u002Fitems`, {\n    headers: { Authorization: `Bearer ${user.token}` },\n    tags: { endpoint: 'list-items' },\n  });\n  check(res, {\n    'status 200':       (r) => r.status === 200,\n    'response \u003C 800ms': (r) => r.timings.duration \u003C 800,\n  });\n  \u002F\u002F Do NOT add sleep() with constant-arrival-rate — pacing is controlled by `rate`\u002F`timeUnit`\n}\n```\n\n## Soak \u002F Endurance Test\n\n```javascript\nexport const options = {\n  stages: [\n    { duration: '5m',  target: 100 }, \u002F\u002F ramp-up\n    { duration: '8h',  target: 100 }, \u002F\u002F endurance plateau (adjust to 3–72h based on requirements)\n    { duration: '5m',  target: 0   }, \u002F\u002F ramp-down\n  ],\n  thresholds: {\n    http_req_failed:   ['rate\u003C0.01'],\n    http_req_duration: [\n      'p(95)\u003C800',\n      { threshold: 'p(99)\u003C2000', abortOnFail: true, delayAbortEval: '60s' },\n    ],\n  },\n};\n```\n\n**Soak test minimum durations (from k6 official guidance)**:\n- Minimum useful: 3–4 hours\n- Standard: 8–12 hours\n- Extended: 24–72 hours\n\n**What to watch for in .NET during soak**:\n- `http_req_duration` p99 trending upward continuously (memory leak \u002F GC pressure)\n- `http_req_failed` rate climbing over time (degrading reliability)\n- `dropped_iterations` counter growing (open model only — indicates saturation)\n- Gen 2 GC collection rate increasing (monitor via Application Insights \u002F dotnet-counters)\n- Thread pool queue length increasing (connection pool exhaustion)\n\n## Multi-Scenario Weighted Load (Realistic Traffic Mix)\n\n```javascript\nimport http from 'k6\u002Fhttp';\n\nexport const options = {\n  scenarios: {\n    \u002F\u002F 70% of load: read endpoint at 70 RPS\n    list_items: {\n      executor: 'constant-arrival-rate',\n      rate: 70, timeUnit: '1s', duration: '10m',\n      preAllocatedVUs: 20, maxVUs: 100,\n      exec: 'listItems',\n      tags: { endpoint: 'list' },\n    },\n    \u002F\u002F 20% of load: write endpoint at 20 RPS\n    create_item: {\n      executor: 'constant-arrival-rate',\n      rate: 20, timeUnit: '1s', duration: '10m',\n      preAllocatedVUs: 10, maxVUs: 50,\n      exec: 'createItem',\n      tags: { endpoint: 'create' },\n    },\n    \u002F\u002F 10% of load: auth endpoint at 10 RPS\n    authenticate: {\n      executor: 'constant-arrival-rate',\n      rate: 10, timeUnit: '1s', duration: '10m',\n      preAllocatedVUs: 5, maxVUs: 20,\n      exec: 'authenticate',\n      tags: { endpoint: 'auth' },\n    },\n  },\n  thresholds: {\n    'http_req_duration{endpoint:list}':     ['p(95)\u003C200'],\n    'http_req_duration{endpoint:create}':   ['p(95)\u003C500'],\n    'http_req_duration{endpoint:auth}':     ['p(95)\u003C300'],\n    'http_req_failed':                       ['rate\u003C0.01'],\n  },\n};\n\nexport function listItems() {\n  http.get(`${__ENV.BASE_URL}\u002Fapi\u002Fitems`);\n}\n\nexport function createItem() {\n  http.post(`${__ENV.BASE_URL}\u002Fapi\u002Fitems`, JSON.stringify({ name: 'test' }), {\n    headers: { 'Content-Type': 'application\u002Fjson' },\n  });\n}\n\nexport function authenticate() {\n  http.post(`${__ENV.BASE_URL}\u002Fapi\u002Fauth\u002Ftoken`,\n    JSON.stringify({ username: 'testuser', password: 'testpass' }),\n    { headers: { 'Content-Type': 'application\u002Fjson' } });\n}\n\nexport default function () { \u002F* not used when scenarios define exec *\u002F }\n```\n\n## Key Load Test Metrics\n\n| Metric | Type | Use For |\n|--------|------|---------|\n| `http_reqs` | Counter | Throughput — `N\u002Fs` in summary = actual RPS |\n| `http_req_duration` | Trend | Primary SLO: check p90\u002Fp95\u002Fp99 |\n| `http_req_waiting` | Trend | TTFB — server-side processing time |\n| `http_req_failed` | Rate | Error rate \u002F error budget |\n| `http_req_blocked` | Trend | TCP connection pool exhaustion signal |\n| `data_sent` \u002F `data_received` | Counter | Bandwidth \u002F payload size |\n| `vus_max` | Gauge | If equals `maxVUs`, executor was constrained |\n| `dropped_iterations` | Counter | Non-zero = system or VU pool couldn't keep up |\n| `checks` | Rate | Functional correctness under load |\n\n**Throughput formula**: `RPS = http_reqs.count \u002F test_duration_seconds`\n\n## SLO Thresholds for .NET REST APIs\n\nStarting point (adjust from production APM data):\n\n```javascript\nthresholds: {\n  http_req_failed:   ['rate\u003C0.01'],           \u002F\u002F 99.0% success SLO\n  http_req_duration: [\n    'p(90)\u003C400',    \u002F\u002F 90th percentile\n    'p(95)\u003C800',    \u002F\u002F 95th percentile (primary SLO)\n    'p(99)\u003C2000',   \u002F\u002F 99th — allows for .NET GC pauses\n  ],\n  http_req_waiting:  ['p(95)\u003C600'],           \u002F\u002F TTFB guard\n  checks:            ['rate>0.99'],           \u002F\u002F functional correctness\n  \u002F\u002F Throughput floor\n  'http_reqs':       ['rate>50'],             \u002F\u002F minimum 50 RPS required\n}\n```\n\n**Error budget guidance**:\n- Mission-critical APIs: \u003C 0.1% error rate\n- Standard APIs: \u003C 1% error rate\n- Health endpoints: p99 \u003C 50ms; error rate \u003C 0.01%\n\n## SharedArray for Large Test Data\n\n```javascript\nimport { SharedArray } from 'k6\u002Fdata';\n\n\u002F\u002F Parsed ONCE — shared read-only across all VUs (critical for large datasets)\nconst users = new SharedArray('test-users', function () {\n  return JSON.parse(open('.\u002Fusers.json'));\n});\n\n\u002F\u002F At 100k rows: SharedArray ≈ 250MB vs non-shared ≈ 9GB\nexport default function () {\n  const user = users[Math.floor(Math.random() * users.length)];\n  http.get(`${__ENV.BASE_URL}\u002Fapi\u002Fprofile\u002F${user.id}`);\n}\n```\n\n**SharedArray rules**:\n- All processing (filtering, mapping) must happen inside the constructor callback\n- Never call `.filter()` \u002F `.map()` on a `SharedArray` outside the constructor\n- Never return a `SharedArray` from `setup()` — causes marshalling, loses benefits\n- Open the file inside the constructor, not at the top level\n\n## GitHub Actions CI\u002FCD Integration\n\n```yaml\nname: Load Tests\non:\n  push:\n    branches: [main]\n  schedule:\n    - cron: '0 2 * * *'  # nightly regression run\n\njobs:\n  load-test:\n    runs-on: ubuntu-latest\n    steps:\n      - uses: actions\u002Fcheckout@v4\n\n      - name: Setup k6\n        uses: grafana\u002Fsetup-k6-action@v1\n        with:\n          k6-version: '0.56.0'   # pin for reproducibility\n\n      - name: Run load tests\n        uses: grafana\u002Frun-k6-action@v1\n        with:\n          path: .\u002Ftests\u002Fload\u002F*.js\n          flags: >-\n            --out json=results\u002Fk6-output.json\n            --env BASE_URL=${{ vars.API_BASE_URL }}\n            --summary-trend-stats=\"avg,min,med,max,p(90),p(95),p(99)\"\n          fail-fast: true\n\n      - name: Upload results\n        if: always()\n        uses: actions\u002Fupload-artifact@v4\n        with:\n          name: k6-load-test-results\n          path: results\u002F\n          retention-days: 30\n```\n\n**Important**: Use `grafana\u002Fsetup-k6-action@v1` + `grafana\u002Frun-k6-action@v1`. Do NOT use the archived `grafana\u002Fk6-action` (deprecated July 2024).\n\n**Exit code behavior**: k6 exits `99` when any threshold fails → GitHub Actions step fails automatically. No extra YAML needed.\n\n## Interpreting the End-of-Test Summary\n\n```\n█ THRESHOLDS\n  http_req_duration\n    ✓ 'p(95)\u003C800'    p(95)=148ms      ← PASSED\n    ✗ 'p(99)\u003C2000'   p(99)=2523ms     ← FAILED → exit 99\n\n█ TOTAL RESULTS\n  http_req_duration...: avg=140ms  min=119ms  med=140ms  max=2523ms  p(90)=180ms  p(95)=148ms  p(99)=2523ms\n  http_req_failed.....: 0.00%\n  http_reqs...........: 5000 8.3\u002Fs         ← actual throughput vs target\n  vus_max.............: 20   min=5   max=20 ← if near maxVUs, executor was constrained\n  dropped_iterations..: 0                   ← non-zero = saturation signal\n```\n\n**Field guide**:\n- `avg >> med` → high-latency outliers present; investigate p99\n- `http_reqs N\u002Fs` → compare to target arrival rate; divergence = saturation\n- `vus_max == maxVUs` → VU pool was constrained; increase `maxVUs` or reduce load\n- `dropped_iterations > 0` → system cannot sustain target rate at `maxVUs`\n- `iteration_duration gap vs http_req_duration` → the gap is think time + overhead\n\n## Sources\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Ftesting-guides\u002Ftest-types\u002Fload-testing\u002F\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Ftesting-guides\u002Ftest-types\u002Fsoak-testing\u002F\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Fusing-k6\u002Fscenarios\u002Fexecutors\u002Fconstant-arrival-rate\u002F\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Fusing-k6\u002Fscenarios\u002Fconcepts\u002Fopen-vs-closed\u002F\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Fjavascript-api\u002Fk6-data\u002Fsharedarray\u002F\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Fresults-output\u002Fend-of-test\u002F\n- https:\u002F\u002Fgithub.com\u002Fgrafana\u002Fsetup-k6-action\n- https:\u002F\u002Fgithub.com\u002Fgrafana\u002Frun-k6-action\n","# k6 Load Testing\n\n## When to Use\nInvoke this skill when the task is to validate that the system sustains its SLOs at **realistic production throughput over time** — including endurance (soak) testing for memory leaks, connection pool exhaustion, and GC pressure.\n\nThis skill covers:\n- **Average-load tests**: validate SLOs at expected normal traffic (use for regression gates)\n- **Soak \u002F endurance tests**: validate reliability under sustained load for hours\n- **Multi-scenario weighted load**: model realistic endpoint traffic mixes\n\nFor above-normal load, use `k6-stress-testing`. For basic smoke\u002Fbaseline validation, use `k6-performance-testing`.\n\n**Always run in this order**: smoke → average-load → stress → soak. Never run a soak test on a system that hasn't first passed average-load and stress.\n\n## Executor Selection\n\n| Executor | Model | When to Use |\n|----------|-------|-------------|\n| `constant-vus` | Closed | Specific concurrent user count; throughput varies with latency |\n| `constant-arrival-rate` | **Open** | Fixed RPS regardless of response time. **Preferred for REST API load tests** |\n| `per-vu-iterations` | Closed | Partition fixed test data across VUs (each VU processes its own slice) |\n\n**Prefer `constant-arrival-rate`** for REST API load tests. In closed models, slow responses reduce RPS invisibly — the system appears healthy while throughput degrades.\n\n## Constant Arrival Rate Load Test (REST API)\n\n```javascript\nimport http from 'k6\u002Fhttp';\nimport { check, sleep } from 'k6';\nimport { SharedArray } from 'k6\u002Fdata';\n\nconst users = new SharedArray('users', function () {\n  return JSON.parse(open('.\u002Ftest-data\u002Fusers.json'));\n});\n\nexport const options = {\n  scenarios: {\n    constant_load: {\n      executor: 'constant-arrival-rate',\n      rate: 50,                 \u002F\u002F 50 iterations per second\n      timeUnit: '1s',\n      duration: '10m',\n      preAllocatedVUs: 20,      \u002F\u002F initial VU pool\n      maxVUs: 100,              \u002F\u002F hard ceiling\n    },\n  },\n  thresholds: {\n    http_req_failed:   ['rate\u003C0.01'],          \u002F\u002F \u003C1% error rate\n    http_req_duration: [\n      'p(90)\u003C400',\n      'p(95)\u003C800',\n      'p(99)\u003C2000',\n    ],\n    http_req_waiting:  ['p(95)\u003C600'],          \u002F\u002F TTFB SLO\n    checks:            ['rate>0.99'],\n    dropped_iterations: ['count\u003C10'],          \u002F\u002F near-zero tolerance for dropped iterations\n  },\n};\n\nexport default function () {\n  const user = users[Math.floor(Math.random() * users.length)];\n  const res = http.get(`${__ENV.BASE_URL}\u002Fapi\u002Fitems`, {\n    headers: { Authorization: `Bearer ${user.token}` },\n    tags: { endpoint: 'list-items' },\n  });\n  check(res, {\n    'status 200':       (r) => r.status === 200,\n    'response \u003C 800ms': (r) => r.timings.duration \u003C 800,\n  });\n  \u002F\u002F Do NOT add sleep() with constant-arrival-rate — pacing is controlled by `rate`\u002F`timeUnit`\n}\n```\n\n## Soak \u002F Endurance Test\n\n```javascript\nexport const options = {\n  stages: [\n    { duration: '5m',  target: 100 }, \u002F\u002F ramp-up\n    { duration: '8h',  target: 100 }, \u002F\u002F endurance plateau (adjust to 3–72h based on requirements)\n    { duration: '5m',  target: 0   }, \u002F\u002F ramp-down\n  ],\n  thresholds: {\n    http_req_failed:   ['rate\u003C0.01'],\n    http_req_duration: [\n      'p(95)\u003C800',\n      { threshold: 'p(99)\u003C2000', abortOnFail: true, delayAbortEval: '60s' },\n    ],\n  },\n};\n```\n\n**Soak test minimum durations (from k6 official guidance)**:\n- Minimum useful: 3–4 hours\n- Standard: 8–12 hours\n- Extended: 24–72 hours\n\n**What to watch for in .NET during soak**:\n- `http_req_duration` p99 trending upward continuously (memory leak \u002F GC pressure)\n- `http_req_failed` rate climbing over time (degrading reliability)\n- `dropped_iterations` counter growing (open model only — indicates saturation)\n- Gen 2 GC collection rate increasing (monitor via Application Insights \u002F dotnet-counters)\n- Thread pool queue length increasing (connection pool exhaustion)\n\n## Multi-Scenario Weighted Load (Realistic Traffic Mix)\n\n```javascript\nimport http from 'k6\u002Fhttp';\n\nexport const options = {\n  scenarios: {\n    \u002F\u002F 70% of load: read endpoint at 70 RPS\n    list_items: {\n      executor: 'constant-arrival-rate',\n      rate: 70, timeUnit: '1s', duration: '10m',\n      preAllocatedVUs: 20, maxVUs: 100,\n      exec: 'listItems',\n      tags: { endpoint: 'list' },\n    },\n    \u002F\u002F 20% of load: write endpoint at 20 RPS\n    create_item: {\n      executor: 'constant-arrival-rate',\n      rate: 20, timeUnit: '1s', duration: '10m',\n      preAllocatedVUs: 10, maxVUs: 50,\n      exec: 'createItem',\n      tags: { endpoint: 'create' },\n    },\n    \u002F\u002F 10% of load: auth endpoint at 10 RPS\n    authenticate: {\n      executor: 'constant-arrival-rate',\n      rate: 10, timeUnit: '1s', duration: '10m',\n      preAllocatedVUs: 5, maxVUs: 20,\n      exec: 'authenticate',\n      tags: { endpoint: 'auth' },\n    },\n  },\n  thresholds: {\n    'http_req_duration{endpoint:list}':     ['p(95)\u003C200'],\n    'http_req_duration{endpoint:create}':   ['p(95)\u003C500'],\n    'http_req_duration{endpoint:auth}':     ['p(95)\u003C300'],\n    'http_req_failed':                       ['rate\u003C0.01'],\n  },\n};\n\nexport function listItems() {\n  http.get(`${__ENV.BASE_URL}\u002Fapi\u002Fitems`);\n}\n\nexport function createItem() {\n  http.post(`${__ENV.BASE_URL}\u002Fapi\u002Fitems`, JSON.stringify({ name: 'test' }), {\n    headers: { 'Content-Type': 'application\u002Fjson' },\n  });\n}\n\nexport function authenticate() {\n  http.post(`${__ENV.BASE_URL}\u002Fapi\u002Fauth\u002Ftoken`,\n    JSON.stringify({ username: 'testuser', password: 'testpass' }),\n    { headers: { 'Content-Type': 'application\u002Fjson' } });\n}\n\nexport default function () { \u002F* not used when scenarios define exec *\u002F }\n```\n\n## Key Load Test Metrics\n\n| Metric | Type | Use For |\n|--------|------|---------|\n| `http_reqs` | Counter | Throughput — `N\u002Fs` in summary = actual RPS |\n| `http_req_duration` | Trend | Primary SLO: check p90\u002Fp95\u002Fp99 |\n| `http_req_waiting` | Trend | TTFB — server-side processing time |\n| `http_req_failed` | Rate | Error rate \u002F error budget |\n| `http_req_blocked` | Trend | TCP connection pool exhaustion signal |\n| `data_sent` \u002F `data_received` | Counter | Bandwidth \u002F payload size |\n| `vus_max` | Gauge | If equals `maxVUs`, executor was constrained |\n| `dropped_iterations` | Counter | Non-zero = system or VU pool couldn't keep up |\n| `checks` | Rate | Functional correctness under load |\n\n**Throughput formula**: `RPS = http_reqs.count \u002F test_duration_seconds`\n\n## SLO Thresholds for .NET REST APIs\n\nStarting point (adjust from production APM data):\n\n```javascript\nthresholds: {\n  http_req_failed:   ['rate\u003C0.01'],           \u002F\u002F 99.0% success SLO\n  http_req_duration: [\n    'p(90)\u003C400',    \u002F\u002F 90th percentile\n    'p(95)\u003C800',    \u002F\u002F 95th percentile (primary SLO)\n    'p(99)\u003C2000',   \u002F\u002F 99th — allows for .NET GC pauses\n  ],\n  http_req_waiting:  ['p(95)\u003C600'],           \u002F\u002F TTFB guard\n  checks:            ['rate>0.99'],           \u002F\u002F functional correctness\n  \u002F\u002F Throughput floor\n  'http_reqs':       ['rate>50'],             \u002F\u002F minimum 50 RPS required\n}\n```\n\n**Error budget guidance**:\n- Mission-critical APIs: \u003C 0.1% error rate\n- Standard APIs: \u003C 1% error rate\n- Health endpoints: p99 \u003C 50ms; error rate \u003C 0.01%\n\n## SharedArray for Large Test Data\n\n```javascript\nimport { SharedArray } from 'k6\u002Fdata';\n\n\u002F\u002F Parsed ONCE — shared read-only across all VUs (critical for large datasets)\nconst users = new SharedArray('test-users', function () {\n  return JSON.parse(open('.\u002Fusers.json'));\n});\n\n\u002F\u002F At 100k rows: SharedArray ≈ 250MB vs non-shared ≈ 9GB\nexport default function () {\n  const user = users[Math.floor(Math.random() * users.length)];\n  http.get(`${__ENV.BASE_URL}\u002Fapi\u002Fprofile\u002F${user.id}`);\n}\n```\n\n**SharedArray rules**:\n- All processing (filtering, mapping) must happen inside the constructor callback\n- Never call `.filter()` \u002F `.map()` on a `SharedArray` outside the constructor\n- Never return a `SharedArray` from `setup()` — causes marshalling, loses benefits\n- Open the file inside the constructor, not at the top level\n\n## GitHub Actions CI\u002FCD Integration\n\n```yaml\nname: Load Tests\non:\n  push:\n    branches: [main]\n  schedule:\n    - cron: '0 2 * * *'  # nightly regression run\n\njobs:\n  load-test:\n    runs-on: ubuntu-latest\n    steps:\n      - uses: actions\u002Fcheckout@v4\n\n      - name: Setup k6\n        uses: grafana\u002Fsetup-k6-action@v1\n        with:\n          k6-version: '0.56.0'   # pin for reproducibility\n\n      - name: Run load tests\n        uses: grafana\u002Frun-k6-action@v1\n        with:\n          path: .\u002Ftests\u002Fload\u002F*.js\n          flags: >-\n            --out json=results\u002Fk6-output.json\n            --env BASE_URL=${{ vars.API_BASE_URL }}\n            --summary-trend-stats=\"avg,min,med,max,p(90),p(95),p(99)\"\n          fail-fast: true\n\n      - name: Upload results\n        if: always()\n        uses: actions\u002Fupload-artifact@v4\n        with:\n          name: k6-load-test-results\n          path: results\u002F\n          retention-days: 30\n```\n\n**Important**: Use `grafana\u002Fsetup-k6-action@v1` + `grafana\u002Frun-k6-action@v1`. Do NOT use the archived `grafana\u002Fk6-action` (deprecated July 2024).\n\n**Exit code behavior**: k6 exits `99` when any threshold fails → GitHub Actions step fails automatically. No extra YAML needed.\n\n## Interpreting the End-of-Test Summary\n\n```\n█ THRESHOLDS\n  http_req_duration\n    ✓ 'p(95)\u003C800'    p(95)=148ms      ← PASSED\n    ✗ 'p(99)\u003C2000'   p(99)=2523ms     ← FAILED → exit 99\n\n█ TOTAL RESULTS\n  http_req_duration...: avg=140ms  min=119ms  med=140ms  max=2523ms  p(90)=180ms  p(95)=148ms  p(99)=2523ms\n  http_req_failed.....: 0.00%\n  http_reqs...........: 5000 8.3\u002Fs         ← actual throughput vs target\n  vus_max.............: 20   min=5   max=20 ← if near maxVUs, executor was constrained\n  dropped_iterations..: 0                   ← non-zero = saturation signal\n```\n\n**Field guide**:\n- `avg >> med` → high-latency outliers present; investigate p99\n- `http_reqs N\u002Fs` → compare to target arrival rate; divergence = saturation\n- `vus_max == maxVUs` → VU pool was constrained; increase `maxVUs` or reduce load\n- `dropped_iterations > 0` → system cannot sustain target rate at `maxVUs`\n- `iteration_duration gap vs http_req_duration` → the gap is think time + overhead\n\n## Sources\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Ftesting-guides\u002Ftest-types\u002Fload-testing\u002F\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Ftesting-guides\u002Ftest-types\u002Fsoak-testing\u002F\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Fusing-k6\u002Fscenarios\u002Fexecutors\u002Fconstant-arrival-rate\u002F\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Fusing-k6\u002Fscenarios\u002Fconcepts\u002Fopen-vs-closed\u002F\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Fjavascript-api\u002Fk6-data\u002Fsharedarray\u002F\n- https:\u002F\u002Fgrafana.com\u002Fdocs\u002Fk6\u002Flatest\u002Fresults-output\u002Fend-of-test\u002F\n- https:\u002F\u002Fgithub.com\u002Fgrafana\u002Fsetup-k6-action\n- https:\u002F\u002Fgithub.com\u002Fgrafana\u002Frun-k6-action\n",{"name":5,"description":7},"main","00b2c68281930728ee390bb16cec32b88d2caac7",".claude\u002Fskills\u002Fk6-load-testing","https:\u002F\u002Fgithub.com\u002Frafsanulhasan\u002FHypermediaEngine","https:\u002F\u002Fgithub.com\u002Frafsanulhasan","\r\n    MCSD || A-CSD || Software Architect\r\n\r\n","2026-06-06 07:47:06",{"source":8,"skillName":5,"commands":55,"prompt":65,"zipAvailable":66},[56,59,62],{"tool":57,"command":58},"npx","npx skills add rafsanulhasan\u002FHypermediaEngine --skill k6-load-testing",{"tool":60,"command":61},"bunx","bunx skills add rafsanulhasan\u002FHypermediaEngine --skill k6-load-testing",{"tool":63,"command":64},"pnpm","pnpm dlx skills add rafsanulhasan\u002FHypermediaEngine --skill k6-load-testing","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Frafsanulhasan\u002FHypermediaEngine\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fk6-load-testing\u002FSKILL.md",true,[],[69,95,114,133,162,184],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":76,"publishedAt":42},"01M6KWEJD606FG3DBBRM0W4DKW","write-tests","Test implementation skill for HypermediaEngine. Takes a confirmed test plan from design-test-cases and produces compilable, runnable xUnit tests following project conventions. Validates coverage quality through mutation testing. Invoked by the sqa-engineer agent after the test plan is approved.","https:\u002F\u002Fgithub.com\u002Frafsanulhasan\u002FHypermediaEngine\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fwrite-tests\u002FSKILL.md",[75],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[77,78,79,28,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,30],"write tests","write","tests","design","test","project conventions","test plan","Bash","plan","quality","xunit","read","edit","coverage","engineer","mutation testing","project","mutation",{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":101,"publishedAt":42},"01M6KWEF2XQ00C1Y6QCK7YHZ1H","write-documentation","Structured documentation writing workflow for HypermediaEngine. Writes or updates README.md files at the repo root and in each project directory (src\u002F, tests\u002F, samples\u002F, etc.) based on current implementation, ADRs, and architecture decisions.","https:\u002F\u002Fgithub.com\u002Frafsanulhasan\u002FHypermediaEngine\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fwrite-documentation\u002FSKILL.md",[],[102,78,103,104,105,106,107,84,108,109,110,111,88,89,112,93,79,113,31,30],"documentation","write documentation","writing","directory","workflow","architecture","src","repo","decisions","README","files","updates",{"skillNo":115,"name":116,"nameZh":6,"description":117,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":118,"categories":119,"keywords":120,"publishedAt":42},"01M6KWEBDXTXAVPQ3BHS3YYYNT","write-adr","Writes a numbered Architecture Decision Record (ADR) for a significant architectural choice in the HypermediaEngine project. Invoked by the software-architect agent whenever a consequential design decision is made or ratified. Saves the ADR to docs\u002Farchitecture\u002Fdecisions\u002F.","https:\u002F\u002Fgithub.com\u002Frafsanulhasan\u002FHypermediaEngine\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fwrite-adr\u002FSKILL.md",[],[121,78,122,80,123,107,124,125,126,127,110,88,128,129,93,130,131,30,132],"adr","write adr","docs","architect","decision","choice","grep","architecture decision record","record","software","opus","architecture decision",{"skillNo":134,"name":135,"nameZh":6,"description":136,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":137,"categories":138,"keywords":145,"publishedAt":42},"01M6KWE76TM3QHGQ08PEB3KY6D","tunit-playwright-ui-testing","Comprehensive guidance for writing coded end-to-end UI tests in C# using TUnit.Playwright. Tests inherit from PageTest — the base class manages all browser lifecycle. Covers NuGet setup, PageTest inheritance, Page Object Model, browser configuration, parallelism limiting, and CI integration. Based on the official TUnit docs at https:\u002F\u002Ftunit.dev\u002Fdocs\u002Fexamples\u002Fplaywright.","https:\u002F\u002Fgithub.com\u002Frafsanulhasan\u002FHypermediaEngine\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Ftunit-playwright-ui-testing\u002FSKILL.md",[139,140],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":141,"categoryName":142,"taxonomyType":143,"taxonomyTypeDesc":144},"01JBSEEDCAT000000000000032","设计师","PROFESSION","适用职业",[28,146,147,148,149,150,151,104,152,153,154,155,156,157,158,159,79,160,161,30],"playwright","ui","playwright ui testing","ui testing","tunit","tunit playwright ui testing","official","dev","examples","ci","lifecycle","model","page","inheritance","end","object",{"skillNo":163,"name":164,"nameZh":6,"description":165,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":166,"categories":167,"keywords":171,"publishedAt":42},"01M6KWE44A738GDK65QS02XJBG","triage","Task classification, decomposition, dependency mapping, and routing skill. Invoked by the triage-agent to break any user request into discrete, independently-routable work items with assigned priorities and agent chains.","https:\u002F\u002Fgithub.com\u002Frafsanulhasan\u002FHypermediaEngine\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Ftriage\u002FSKILL.md",[168],{"categoryNo":169,"categoryName":170,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发",[164,172,173,174,175,176,36,177,178,179,180,181,182,183],"routing","classification","work","user","dependency","priorities","decomposition","user request","mapping","with","break","request",{"skillNo":185,"name":186,"nameZh":6,"description":187,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":188,"categories":189,"keywords":194,"publishedAt":42},"01M6KWDZ3W9ZKPR7FTTTSX8H2K","system-design","Expert system design skill. Use PROACTIVELY to design systems, review architectures, and assist software engineers in learning design principles and implementing them in the code with design patterns.","https:\u002F\u002Fgithub.com\u002Frafsanulhasan\u002FHypermediaEngine\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fsystem-design\u002FSKILL.md",[190,193],{"categoryNo":191,"categoryName":192,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000015","视觉设计",{"categoryNo":141,"categoryName":142,"taxonomyType":143,"taxonomyTypeDesc":144},[80,38,195,196,197,198,199,127,88,200,201,202,203,204,205,130,206,207,181,208],"system design","web","design systems","learning","haiku","use","design patterns","design principles","expert","patterns","systems","architectures","principles","implementing"]