2nd/10_Wiki/Topics/AI_and_ML/Modern Engineering Practices (현대적 엔지니어링 프랙티스).md

---
id: wiki-2026-0508-modern-engineering-practices-현대적
title: "Modern Engineering Practices (현대적 엔지니어링 프랙티스)"
category: 10_Wiki/Topics
status: verified
canonical_id: self
aliases: [Modern Engineering, Modern Dev Practices, 현대 엔지니어링, Engineering Best Practices 2026]
duplicate_of: none
source_trust_level: A
confidence_score: 0.92
verification_status: applied
tags: [engineering, devops, ci-cd, dora, sre, observability, ai-augmented, trunk-based, iac]
raw_sources: []
last_reinforced: 2026-05-10
github_commit: pending
tech_stack: { language: any, framework: github-actions|terraform|opentelemetry }
---

# Modern Engineering Practices (현대적 엔지니어링 프랙티스)

> 한 줄: 2026년 표준 — 작은 PR·trunk-based·CI/CD·IaC·observability·DORA 측정·AI-augmented dev. "고객까지 30분 안에 배포".

## 핵심
- **Flow**: trunk-based + short-lived branch (≤1일), feature flag, 작은 PR (<400 LOC), 즉시 review, 즉시 merge.
- **CI/CD**: 모든 PR에서 build·test·lint·security scan, 머지 즉시 staging, canary/progressive rollout (LaunchDarkly·Argo Rollouts).
- **DORA 4 metrics**: deployment frequency, lead time for changes, change failure rate, MTTR. Elite = 다회/일·<1h·<15%·<1h.
- **IaC**: Terraform/OpenTofu·Pulumi·Crossplane. Drift detection. GitOps (Argo CD, Flux).
- **Observability**: 3 pillars (logs·metrics·traces) + profiles. OpenTelemetry 표준, Grafana/Datadog/Honeycomb.
- **AI-augmented**: Copilot/Cursor/Claude Code 코드 작성·PR review·테스트 생성. SLSA·SBOM 등 supply-chain security.

## 결정 기준
| 영역 | 채택 | 회피 |
|---|---|---|
| 브랜치 전략 | trunk-based + short branch + flag | git-flow (대형 long-lived branch) |
| 배포 | progressive (canary 1%→10%→100%) | big-bang weekend deploy |
| 테스트 피라미드 | unit > contract > integ > e2e | 역피라미드 (UI 무거움) |
| Secret 관리 | Vault, SOPS, cloud KMS | env 파일 git 커밋 |
| Infra | IaC + GitOps | 콘솔 클릭 |
| 모니터링 | OTel + SLO 알람 | 모든 에러 페이지 |
| Code review | < 4시간 응답, 작은 PR | 1000줄 PR |

## 💻 패턴

### GitHub Actions: PR + main 분리
```yaml
name: ci
on:
  pull_request:
  push: { branches: [main] }
jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-node@v4
        with: { node-version: 20, cache: pnpm }
      - run: pnpm install --frozen-lockfile
      - run: pnpm lint && pnpm test --coverage
      - uses: codecov/codecov-action@v4
  deploy:
    needs: test
    if: github.ref == 'refs/heads/main'
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - run: ./scripts/deploy.sh staging
```

### Trunk-based + feature flag
```ts
import { useFlag } from "@/lib/flags";
export function Checkout() {
  const newFlow = useFlag("checkout-v2", { default: false });
  return newFlow ? <CheckoutV2 /> : <CheckoutV1 />;
}
```

### Terraform 모듈 + remote state
```hcl
terraform {
  required_version = ">= 1.7"
  backend "s3" {
    bucket = "tfstate-prod" key = "app/terraform.tfstate"
    region = "us-east-1" dynamodb_table = "tf-locks"
  }
}
module "api" {
  source  = "./modules/service"
  name    = "api" image = "ghcr.io/org/api:${var.git_sha}"
  min_replicas = 3 cpu = "500m"
}
```

### OpenTelemetry instrumentation (Node)
```ts
import { NodeSDK } from "@opentelemetry/sdk-node";
import { OTLPTraceExporter } from "@opentelemetry/exporter-trace-otlp-http";
import { getNodeAutoInstrumentations } from "@opentelemetry/auto-instrumentations-node";

new NodeSDK({
  traceExporter: new OTLPTraceExporter({ url: process.env.OTEL_ENDPOINT }),
  instrumentations: [getNodeAutoInstrumentations()],
}).start();
```

### Progressive delivery (Argo Rollouts)
```yaml
apiVersion: argoproj.io/v1alpha1
kind: Rollout
spec:
  strategy:
    canary:
      steps:
        - setWeight: 5
        - pause: { duration: 5m }
        - setWeight: 25
        - analysis: { templates: [{ templateName: success-rate }] }
        - setWeight: 100
```

### SLO 정의 (Prometheus)
```yaml
- alert: APILatencyBudgetBurn
  expr: (1 - histogram_quantile(0.99, rate(http_request_duration_seconds_bucket[5m]))/0.3) < 0.5
  for: 10m
  annotations: { summary: "P99 latency burn rate" }
```

### Pre-commit + secret scan
```yaml
# .pre-commit-config.yaml
repos:
  - repo: https://github.com/gitleaks/gitleaks
    rev: v8.21.0
    hooks: [{ id: gitleaks }]
  - repo: https://github.com/pre-commit/pre-commit-hooks
    rev: v5.0.0
    hooks: [{ id: trailing-whitespace }, { id: end-of-file-fixer }]
```

### AI-augmented PR review (Claude/Copilot)
```yaml
# GitHub Action에서 PR diff → LLM review
- uses: anthropics/claude-code-action@v1
  with:
    anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}
    prompt: "Review for bugs/security/perf, suggest tests"
```

## 🔗 Graph
- 상위: [[DevOps]] · [[SRE]]
- 관련: [[CI CD]] · [[Trunk-Based-Development]] · [[GitOps]] · [[Terraform]] · [[OpenTelemetry]] · [[DORA]] · [[Feature-Flags]] · [[Progressive-Delivery]] · [[SLSA]] · [[Code-Review]]
- AI 결합: [[Cursor]] · [[Claude-Code]]

## 🤖 LLM 활용
- PR 리뷰 1차 자동화, 테스트 생성, 마이그레이션 스크립트, 문서 초안.
- runbook 생성: 알람 → LLM이 과거 인시던트 컨텍스트로 대응 가이드.
- 위험: 보안·라이선스 검증 없이 머지 금지. 사람 review 필수.

## ❌ 안티패턴
- **DORA 측정 없는 "DevOps"** — 베이스라인 없이 개선 불가.
- **Long-lived feature branch** — merge hell. trunk + flag.
- **모든 알람 = 페이지** — alert fatigue. SLO burn rate 알람만 페이지.
- **수동 prod 배포** — 재현·롤백 불가. CI/CD 강제.
- **Secret을 env 파일 커밋** — Vault·KMS·SOPS.
- **Test 커버리지만 KPI** — gaming. mutation testing·핵심 path 우선.
- **AI 코드 무검토 머지** — hallucination·라이선스·보안 이슈. 동일한 review 게이트.

## 🧪 검증 / 중복
- 중복 후보: [[CI CD]], [[DevOps]], [[SRE]] — 본 문서는 우산 (실천 종합), 각 페이지는 deep dive.
- 검증: DORA 자동 수집 (Sleuth, LinearB, Faros) · SLO 대시보드 · CFR < 15%.

## 🕓 Changelog
- 2026-05-08 | Phase 1 — 자동 시드.
- 2026-05-10 | Manual cleanup — DORA·trunk-based·Terraform·OTel·Argo Rollouts·AI-augmented 패턴 정리.