Files
2nd/10_Wiki/Topics/Architecture/소프트웨어 아키텍처 베스트 프랙티스.md
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

4.7 KiB

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
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
wiki-2026-0508-소프트웨어-아키텍처-베스트-프랙티스 소프트웨어 아키텍처 베스트 프랙티스 10_Wiki/Topics verified self
Architecture Best Practices
SW Architecture Patterns
none A 0.9 applied
architecture
best-practices
patterns
2026-05-10 pending
language framework
agnostic agnostic

소프트웨어 아키텍처 베스트 프랙티스

매 한 줄

"매 best practice는 매 context-bound이며, 매 trade-off 의 명시 없이 적용 시 매 cargo cult로 전락한다". Fowler / Vernon / Newman의 매 distillation은 매 boundary 의 명확화, 매 dependency 의 inversion, 매 deployment 의 independence를 매 핵심으로 본다.

매 핵심

매 SOLID at architecture level

  • SRP: 매 service 매 single business capability.
  • OCP: 매 plugin / 매 extension point.
  • LSP: 매 contract-based versioning.
  • ISP: 매 client-specific API (BFF).
  • DIP: 매 ports & adapters.

매 Cloud-native principles (12-factor + extensions)

  • 매 stateless processes.
  • 매 config via env.
  • 매 backing services as attached resources.
  • 매 disposability (fast startup, graceful shutdown).
  • 매 logs as event streams.
  • 매 health endpoints (/healthz, /readyz).
  • 매 telemetry by default (OpenTelemetry).

매 응용

  1. 매 deployment 독립성으로 매 release cadence 향상.
  2. 매 boundary 명확화로 매 team autonomy.
  3. 매 observability 매 ship-time guarantee.

💻 패턴

Health endpoint (Spring Boot)

@RestController
class HealthController {
    @GetMapping("/healthz") // liveness
    ResponseEntity<?> live() { return ResponseEntity.ok().build(); }

    @GetMapping("/readyz") // readiness
    ResponseEntity<?> ready(DataSource ds) {
        try (var c = ds.getConnection()) {
            c.prepareStatement("SELECT 1").execute();
            return ResponseEntity.ok().build();
        } catch (Exception e) {
            return ResponseEntity.status(503).build();
        }
    }
}

Circuit breaker (Resilience4j)

CircuitBreaker cb = CircuitBreaker.of("payment",
    CircuitBreakerConfig.custom()
        .failureRateThreshold(50)
        .slowCallDurationThreshold(Duration.ofSeconds(2))
        .waitDurationInOpenState(Duration.ofSeconds(30))
        .build());

Supplier<Receipt> guarded = CircuitBreaker
    .decorateSupplier(cb, () -> paymentClient.charge(req));
return Try.ofSupplier(guarded).recover(t -> Receipt.queued()).get();

Outbox pattern (Postgres + Debezium)

BEGIN;
INSERT INTO orders (id, status) VALUES ($1, 'CONFIRMED');
INSERT INTO outbox (aggregate_id, type, payload)
  VALUES ($1, 'OrderConfirmed', $2);
COMMIT;
-- Debezium reads outbox table → publishes to Kafka

Idempotency key (REST)

@app.post("/payments")
def charge(req: ChargeReq, idempotency_key: str = Header(...)):
    cached = redis.get(f"idem:{idempotency_key}")
    if cached:
        return json.loads(cached)
    result = process(req)
    redis.setex(f"idem:{idempotency_key}", 86400, json.dumps(result))
    return result

Feature flag (LaunchDarkly-style)

if (await flags.isEnabled("new-checkout", { userId })) {
  return newCheckout(cart);
}
return legacyCheckout(cart);

매 결정 기준

상황 Approach
매 cross-service consistency Outbox + Kafka
매 retry safety Idempotency key
매 cascading failure 방지 Circuit breaker
매 zero-downtime release Feature flag + blue/green
매 schema evolution Backward-compatible (additive) only

기본값: 매 12-factor + OpenTelemetry + idempotency.

🔗 Graph

🤖 LLM 활용

언제: 매 best practice 의 trade-off 설명 / 매 idempotency key generation. 언제 X: 매 specific SLA tuning — 매 production data 기반 결정 필요.

안티패턴

  • Cargo cult microservices: 매 boundary 없이 매 split.
  • Distributed transaction: 매 2PC across services.
  • Shared database: 매 service 간 매 schema 공유.
  • Dual write: 매 outbox 없이 매 DB + Kafka 동시 write.

🧪 검증 / 중복

  • Verified (Newman, Building Microservices 2e; Vernon, Implementing DDD; Fowler, Patterns of Enterprise Application Architecture).
  • 신뢰도 A.

🕓 Changelog

날짜 변경
2026-05-08 Phase 1
2026-05-10 Manual cleanup — 매 outbox / circuit breaker / idempotency 추가