History

nns 019c57ae3b Some checks are pending Auto-tag / Create date-tag (push) Waiting to run Details CI / Backend (.NET 8) (push) Waiting to run Details CI / Web (React + Vite) (push) Waiting to run Details CI / POS (WPF, Windows) (push) Waiting to run Details Docker API / Build + push API (push) Waiting to run Details Docker API / Deploy API on stage (push) Blocked by required conditions Details feat(s25): autonomous continuous quality monitoring (8/8) Hourly smoke watchdog + auto-fix loop + dashboard + multi-tenant guard + perf regression + cleanup job + README badge. 1. ~/quality-watchdog.sh (cron 5 * * * ) — 8 checks (~60s): /health/ready, signup→login→/api/me, GET products, Playwright UI smoke (3.1 product CRUD), /metrics format, /hubs/notifications negotiate with token, multi-tenant isolation, perf p95. 2. Auto-fix loop: 2× consecutive red → ~/.fm-watchdog/incident-.txt + queue/0000-incident-* to bump it ahead of Server-Claude's sprint queue. fm-watchdog.sh sees prefix 0000- as next. 3. scripts/quality-dashboard.py — renders docs/quality-status.md (current emoji, 8-step table, perf baseline, 7-day history, 24-run sparkline) + injects README badge 🟢/🟡/🔴. 4. Multi-tenant smoke: signup 2 orgs `quality-{epoch}-A/B`, create product in A, verify B sees 404/403 + total=0. 5. Perf regression: p95 over 10 reqs for /api/me, products, sales/retail/stats. Baseline = median of last 10 samples (robust to noise). >50% from baseline → alert. First 5 runs always green (warm-up). 6. HousekeepingJobs.PruneQualityTestOrgsAsync (cron 30 2 * * * UTC): finds orgs `quality-%` older than 24h, dynamically scans information_schema for tables with OrganizationId, iteratively DELETEs with FK-violation retry (up to 10 passes), then cleans AspNetUser/OpenIddict by email pattern `quality-%@test-fm.local`, finally users + organizations. 7. README badge: <!-- quality-badge --> marker updated each run. Validated: stage deploy ✓, Hangfire job registered ✓, dry-run SQL on 24 stage candidates → 0 remaining ✓, 3 cron-triggered runs all 8/8 green (12:42/12:45/12:48 +05) ✓. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>		2026-06-08 12:50:35 +05:00
..
lib	test(ui-deep): items 6-9 — Supply/RetailSale/InventoryDocs/Reports	2026-05-30 13:37:01 +05:00
reports	fix(s23): adversarial bug-hunt — 4 bugs found, all fixed	2026-06-08 01:35:50 +05:00
scenarios	test(s15): axe a11y + focus traps + unit coverage 80% + property tests + backup drill	2026-06-07 14:53:38 +05:00
scripts	feat(s25): autonomous continuous quality monitoring (8/8)	2026-06-08 12:50:35 +05:00
.gitignore	feat(e2e): infrastructure + first full-cycle scenario + baseline report	2026-05-08 00:05:52 +05:00
package.json	test(s15): axe a11y + focus traps + unit coverage 80% + property tests + backup drill	2026-06-07 14:53:38 +05:00
playwright.config.ts	fix(stage-tests): IP-limit 60/min, locale ru-RU в playwright, исправлены payload'ы verify-spec'ов	2026-06-04 17:38:36 +05:00
pnpm-lock.yaml	test(s15): axe a11y + focus traps + unit coverage 80% + property tests + backup drill	2026-06-07 14:53:38 +05:00
README.md	feat(e2e): infrastructure + first full-cycle scenario + baseline report	2026-05-08 00:05:52 +05:00
run.sh	feat(e2e): infrastructure + first full-cycle scenario + baseline report	2026-05-08 00:05:52 +05:00
runner.ts	feat(e2e): infrastructure + first full-cycle scenario + baseline report	2026-05-08 00:05:52 +05:00
s23-setup.ts	feat(s25): autonomous continuous quality monitoring (8/8)	2026-06-08 12:50:35 +05:00
tsconfig.json	feat(e2e): infrastructure + first full-cycle scenario + baseline report	2026-05-08 00:05:52 +05:00

README.md

tests/e2e

Декларативные end-to-end сценарии. Один YAML описывает шаги, TypeScript-handler — конкретные API/UI/DB-проверки. Отчёт в Markdown.

Запуск

tests/e2e/run.sh full-cycle              # полный прогон (API + UI)
tests/e2e/run.sh full-cycle --api-only   # без Playwright, только axios + DB

Первый запуск установит node_modules/ (axios, pg, playwright, js-yaml, tsx).

Отчёт: tests/e2e/reports/<scenario>-<timestamp>.md.

Структура

tests/e2e/
├── runner.ts              # entry: парсит YAML, прогоняет steps, пишет report
├── run.sh                 # wrapper: pnpm install + tsx
├── lib/
│   ├── api.ts             # axios + login()
│   ├── db.ts              # docker exec psql, resetTenantData(), countRows()
│   └── report.ts          # markdown-аккумулятор
└── scenarios/
    ├── full-cycle.yml          # декларация шагов
    └── full-cycle.steps.ts     # код handler'ов

Preconditions

Поле reset_db: true в YAML вызывает resetTenantData() в lib/db.ts. Что чистится:

Tenant-таблицы (organizations, employees, supplies, retail_sales, …) — TRUNCATE … CASCADE.
AspNetUsers / users / AspNetUserRoles — оставляем только admin@food-market.local.
OpenIddict tokens — все valid → revoked.

Что НЕ чистится (берегём как baseline):

Реестр товаров: products, product_groups, units_of_measure, product_packagings, product_barcodes, product_prices, product_images.
Системные справочники: countries, currencies, price_types, employee_roles.
__EFMigrationsHistory, OpenIddict* таблицы (структура), platform_settings, system_settings.

После TRUNCATE — smoke login(admin@food-market.local). Если падает — runner выходит с кодом 3.

Добавление сценария

scenarios/<name>.yml — мета + список steps[].id.
scenarios/<name>.steps.ts — экспортируй функции с теми же id (async function stepXX_foo({ ctx, step, report })).
Запусти tests/e2e/run.sh <name>.

ctx — общий объект между шагами для прокидывания id'ов созданных сущностей и токенов сессий.

step.checks — массив проверок (api/ui/db). Если хоть одна ok=false — шаг помечен как fail.

report.bug() / report.ux() / report.gap() / report.perf() — категоризованные находки в финальной секции.

Зависимости

Postgres контейнер food-market-postgres (psql вызывается через docker exec).
API на https://admin.food-market.kz (или E2E_ADMIN_URL env).
Playwright headless chromium для UI-проверок (pnpm exec playwright install chromium если нет).

README.md Unescape Escape