Denozordec
2c92e78b24
quality / commitlint (push) Skipped
CD / update-wiki (push) Successful in 8s
quality / changes (push) Successful in 5s
quality / docker-check (push) Skipped
quality / web (push) Successful in 51s
quality / api (push) Successful in 44s
CD / quality (push) Successful in 1m44s
CD / publish (push) Successful in 1m35s
- Added new health worker URL and token fields to the AppConfig interface, allowing for Cloudflare Worker integration. - Updated health check routes to utilize the new worker configuration, enabling dynamic health checks via Cloudflare Workers. - Introduced a health log endpoint for services, providing detailed logs of health probe results. - Enhanced health check service logic to support both local and Cloudflare Worker providers, improving flexibility in health monitoring. - Updated UI components to reflect changes in health check provider settings and display relevant health information. This commit significantly improves the health check management capabilities, allowing for better integration with Cloudflare Workers and enhanced logging features.
Cloudflare Domain Manager
Self-hosted service for managing domains, DNS records, and SSL certificates via Cloudflare API.
Features
- Domain and DNS record management (hybrid sync with Cloudflare)
- Domain groups (local / vpn / external) and service tags
- Service groups with common FQDN and DNS-based load balancing (Round Robin / Failover / Weighted)
- Health checks (TCP / HTTP) with automatic DNS reconciliation on status change
- SSL certificate expiry monitoring
- React UI with TanStack Router & Query
- Single Docker container deployment
Quick start
cp .env.example .env
# Edit CLOUDFLARE_API_TOKEN, JWT_SECRET
docker compose up -d
Open http://localhost:8080 — default login admin / admin (dev only).
Stack
- Backend: Node.js 22, Fastify 5, Drizzle ORM, SQLite (
apps/api) - Shared: Zod schemas (
packages/shared), DB layer (packages/db) - Frontend: pnpm monorepo (
apps/web+packages/ui), React, Vite, TanStack Router/Query/Table, shadcn/ui (base-nova), Recharts - CI: Gitea Actions (docs/releasing.md) — quality gates + semantic-release + bake (
cfdm/cloudflare-domain-manager)
Development
pnpm install
pnpm --filter @cfdm/shared build
pnpm --filter @cfdm/db build
# API + SQLite (:8080)
pnpm --filter @cfdm/api dev
# Frontend (:5173, proxies /api → :8080)
pnpm --filter web dev
pnpm turbo build
pnpm turbo test
shadcn CLI: cd apps/web && pnpm dlx shadcn@latest add <component>
Documentation
See docs/Home.md, docs/releasing.md and CONTRIBUTING.md.
Load balancing & health checks
Группа сервисов может иметь общий домен (service_groups.domain). На общем домене и на
привязках сервиса с несколькими A-записями включается балансировка и health-check:
- Режимы LB:
round_robin(по одной A на каждый up-IP),failover(A только для up-IP с минимальным приоритетом; бэкапы подключаются при падении primary),weighted(веса учитываются в БД; в Cloudflare free отображается как Round Robin, т.к. CF API не допускает дубликаты(name, type, content)A-записей — для истинного weighted нужен CF Load Balancer). - Health-check: TCP connect или HTTP (настраиваемый порт, путь, ожидаемый статус,
интервал, таймаут). Результаты хранятся в
ip_health_status(up/degraded/down/unknown) и опрашиваются UI с polling 10с. - Reconcile: при смене статуса IP cron перезаписывает A-записи в Cloudflare, оставляя
только активные по политике LB. Реакция = TTL A-записи (
ttl=1proxied — минимальный).
Env для health-check:
| Variable | Default | Description |
|---|---|---|
HEALTH_CHECK_CRON |
*/30 * * * * * |
Cron выражение для запуска проверок (каждые 30с) |
HEALTH_DEGRADED_FAILURES |
1 |
Порог последовательных ошибок → статус degraded |
HEALTH_DOWN_FAILURES |
2 |
Порог последовательных ошибок → статус down (IP убирается из DNS) |
HEALTH_LATENCY_WARN_MS |
1000 |
Латентность выше порога → degraded даже при успешном connect |