Files

T

Antigravity Agent f8b21af4be Wiki cleanup: error-doc removal, dedup merge, link normalization

10_Wiki/Topics 대규모 정리:
- 오류 캡처/미완성 stub 문서 227개 제거
- 교차폴더 중복 43클러스터 병합 (63파일 → redirect)
- 링크명 정규화: 깨진 링크 수정·redirect 직결·개념 매핑 ~2,400건
- 카테고리 MOC 6개 신규 생성
- Graph 섹션 미해결 related-keyword 링크 10,058건 제거

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-05-20 23:52:15 +09:00

6.3 KiB

Raw Blame History

id, title, category, status, canonical_id, aliases, duplicate_of, source_trust_level, confidence_score, verification_status, tags, raw_sources, last_reinforced, github_commit, tech_stack

title

Enterprise Scale Monorepo Management

매 한 줄

"매 single repo, many projects — 매 build graph 의 일관성". 매 monorepo 의 핵심은 모든 코드를 하나의 VCS root 안에 두되, 매 build 시스템이 dependency graph 를 이해해서 affected projects 만 build/test 한다는 점. Google (Piper/Bazel), Meta (Buck2), Microsoft (Rush), 그리고 매 OSS 의 Nx/Turborepo 가 이 패러다임을 driving.

매 핵심

매 Monorepo 가치

Atomic cross-project changes: 매 API + caller 의 한 PR 안에서 변경.
Shared tooling: 매 lint, format, build, test 의 unified config.
Visibility: 매 모든 코드 의 grep-able.
Refactor confidence: 매 type-checker 가 모든 caller 를 검증.

매 도전 과제

Build time scaling: 매 naive build 의 N² growth — affected detection 필수.
VCS performance: 매 git 의 100GB+ repo 에서 sparse checkout / VFS 필요.
Permissions: 매 single repo + multiple teams = CODEOWNERS / branch protection.
CI cost: 매 모든 commit 의 모든 project rebuild 의 X — incremental + cache.

매 도구 선택

Nx (TypeScript-heavy, mid-large): smart caching + computation graph.
Turborepo (Vercel, JS/TS): Rust-based, simple config, remote cache.
Bazel (polyglot, mega-scale): hermetic builds, network-distributed.
Pants (Python-heavy): similar to Bazel, lighter setup.
Rush (.NET / TS): Microsoft's pnpm-based.

💻 패턴

Nx workspace structure

my-org/
├── nx.json
├── package.json
├── tsconfig.base.json
├── apps/
│   ├── web/        # Next.js app
│   └── api/        # NestJS API
├── libs/
│   ├── ui/         # shared React components
│   ├── data-access/# API clients
│   └── utils/
└── tools/

nx.json with affected + cache

{
  "tasksRunnerOptions": {
    "default": {
      "runner": "nx-cloud",
      "options": {
        "cacheableOperations": ["build", "test", "lint", "e2e"],
        "accessToken": "${NX_CLOUD_TOKEN}"
      }
    }
  },
  "targetDefaults": {
    "build": { "dependsOn": ["^build"], "inputs": ["production", "^production"] },
    "test":  { "inputs": ["default", "^production", "{workspaceRoot}/jest.preset.js"] }
  },
  "namedInputs": {
    "default":    ["{projectRoot}/**/*", "sharedGlobals"],
    "production": ["default", "!{projectRoot}/**/*.spec.ts", "!{projectRoot}/jest.config.ts"]
  }
}

Affected detection in CI

# .github/workflows/ci.yml
jobs:
  affected:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with: { fetch-depth: 0 }
      - uses: nrwl/nx-set-shas@v4
      - run: pnpm install --frozen-lockfile
      - run: npx nx affected -t lint test build --parallel=4
      - run: npx nx affected -t e2e --parallel=1

Turborepo pipeline

// turbo.json
{
  "$schema": "https://turbo.build/schema.json",
  "globalDependencies": ["**/.env.*local"],
  "tasks": {
    "build": {
      "dependsOn": ["^build"],
      "outputs": [".next/**", "!.next/cache/**", "dist/**"]
    },
    "test": { "dependsOn": ["build"], "outputs": ["coverage/**"] },
    "lint": {},
    "dev":  { "cache": false, "persistent": true }
  }
}

Bazel BUILD file (polyglot)

# libs/ui/BUILD.bazel
load("@npm//:defs.bzl", "npm_link_all_packages")
load("@aspect_rules_ts//ts:defs.bzl", "ts_project")

ts_project(
    name = "ui",
    srcs = glob(["src/**/*.ts", "src/**/*.tsx"]),
    declaration = True,
    tsconfig = "//:tsconfig",
    deps = [
        "//libs/utils",
        "@npm//react",
        "@npm//@types/react",
    ],
    visibility = ["//apps:__subpackages__"],
)

CODEOWNERS for team boundaries

# CODEOWNERS
/apps/web/        @org/frontend-team
/apps/api/        @org/backend-team
/libs/ui/         @org/design-system
/libs/data-access/ @org/backend-team @org/frontend-team
/.github/         @org/platform-team
/tools/           @org/platform-team

Remote cache with Turborepo

# Self-hosted with turborepo-remote-cache
docker run -p 3000:3000 \
  -e TURBO_TOKEN=secret \
  -e STORAGE_PROVIDER=s3 \
  -e STORAGE_PATH=my-bucket \
  ducktors/turborepo-remote-cache

# In repo:
echo 'TURBO_API=http://cache.internal:3000' > .turbo/config.json
echo 'TURBO_TOKEN=secret' >> .turbo/config.json
turbo build --remote-only

매 결정 기준

상황	Approach
<50 packages, JS/TS only	Turborepo (simplicity)
50-500 packages, type-heavy	Nx (project graph + generators)
Polyglot (Go+Rust+TS+Python)	Bazel (hermeticity)
Python ML monorepo	Pants v2
<10 packages	pnpm workspaces alone

기본값: 매 TS/JS team 에게 Turborepo 또는 Nx (size dependent).

🔗 Graph

부모: Version Control
응용: Code Ownership
Adjacent: Bazel · Nx · Turborepo

🤖 LLM 활용

언제: refactoring across packages, generating new lib boilerplate (Nx generator), CODEOWNERS automation, dependency graph 분석. 언제 X: 매 build cache invalidation logic 의 직접 결정 — Nx/Bazel hash algorithm 신뢰.

❌ 안티패턴

Single CI job for entire repo: 매 affected detection 없이 매 commit 모든 build — 30분+ pipeline.
No remote cache: 매 each developer 가 cold rebuild — wasteful.
Mixing app-specific and lib code: 매 libs/ 와 apps/ 의 분리 안 함 — circular deps risk.
Implicit dependencies: 매 package.json 에 list 안 된 import — Bazel/Nx 가 catch.
No CODEOWNERS: 매 review fatigue + ownership 모호.

🧪 검증 / 중복

Verified (Nx Cloud docs 2026, Turborepo 2.0, Bazel 7).
신뢰도 A.

🕓 Changelog

날짜	변경
2026-05-08	Phase 1
2026-05-10	Manual cleanup — monorepo tooling matrix + Nx/Turbo/Bazel patterns

6.3 KiB Raw Blame History