Files

T

Antigravity Agent 504fd5fb42 [G1-Sync] Manual knowledge update

2026-05-10 22:08:15 +09:00

5.6 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

Architecture Diagramming Standards

매 한 줄

"매 diagram 의 audience 의 결정한다". 매 C4 model (Simon Brown, 2018) 의 Context → Container → Component → Code 의 4-level zoom 의 default standard 의 됨. 매 2026 의 PlantUML + Mermaid + Structurizr DSL 의 most-used toolchain.

매 핵심

매 C4 Model levels

Level 1 — System Context: 매 system + users + external systems. 매 non-technical audience.
Level 2 — Container: 매 deployable units (web app, DB, queue). 매 technical overview.
Level 3 — Component: 매 container 안의 logical components. 매 dev team.
Level 4 — Code: 매 class diagram (rarely used — 매 IDE 의 generation 의 default).

매 supplementary diagrams

Sequence: 매 request flow / interaction.
Deployment: 매 infra topology (k8s, regions).
Dynamic: 매 runtime view, message sequence.
State: 매 entity lifecycle.

매 응용

ADR 의 embedded diagram — 매 context.
Onboarding 의 system overview.
Incident postmortem 의 affected component highlight.

💻 패턴

C4 Context (Mermaid)

C4Context
  title System Context for Banking App
  Person(customer, "Customer", "Bank customer with accounts")
  System(banking, "Internet Banking", "Allows customers to view info")
  System_Ext(mainframe, "Mainframe", "Stores account, txn info")
  System_Ext(email, "Email System", "SMTP relay")
  Rel(customer, banking, "Uses", "HTTPS")
  Rel(banking, mainframe, "Reads/writes", "XML/HTTPS")
  Rel(banking, email, "Sends emails", "SMTP")

C4 Container (PlantUML + C4-PlantUML)

@startuml
!include <C4/C4_Container>
Person(user, "User")
System_Boundary(c1, "Banking System") {
  Container(spa, "SPA", "React/TS", "UI")
  Container(api, "API", "Go/Gin", "JSON/HTTPS")
  ContainerDb(db, "Database", "Postgres 17", "Customers, accounts")
  Container(worker, "Worker", "Go", "Async jobs")
  ContainerQueue(mq, "Queue", "NATS JetStream")
}
Rel(user, spa, "Uses", "HTTPS")
Rel(spa, api, "Calls", "JSON/HTTPS")
Rel(api, db, "Reads/writes")
Rel(api, mq, "Publishes")
Rel(worker, mq, "Consumes")
@enduml

Structurizr DSL (workspace-as-code)

workspace "Banking" {
  model {
    user = person "Customer"
    bank = softwareSystem "Banking" {
      spa = container "SPA" "React/TS"
      api = container "API" "Go/Gin"
      db = container "Database" "Postgres 17" "Database"
    }
    user -> spa "Uses"
    spa -> api "JSON/HTTPS"
    api -> db "SQL"
  }
  views {
    container bank "Containers" { include * autolayout lr }
    theme default
  }
}

Sequence (Mermaid)

sequenceDiagram
  actor U as User
  participant S as SPA
  participant A as API
  participant D as DB
  U->>S: Click "Pay"
  S->>A: POST /payments
  A->>D: BEGIN; INSERT; COMMIT
  A-->>S: 201 Created
  S-->>U: Success toast

Deployment (Mermaid)

flowchart LR
  subgraph k8s[Kubernetes / us-east-1]
    api[API Pods x3]
    worker[Worker Pods x2]
  end
  subgraph rds[RDS]
    pg[(Postgres 17 Multi-AZ)]
  end
  ALB --> api
  api --> pg
  worker --> pg
  worker --> SQS

Diagrams as code (Python diagrams lib)

from diagrams import Diagram, Cluster
from diagrams.aws.compute import ECS
from diagrams.aws.database import RDS
from diagrams.aws.network import ELB

with Diagram("Web", show=False, direction="LR"):
    lb = ELB("lb")
    with Cluster("Services"):
        svc = [ECS("svc1"), ECS("svc2")]
    db = RDS("primary")
    lb >> svc >> db

Auto-generated from infra (Terraform → diagram)

# Inframap / Pluralith / Cloudcraft import
terraform graph | dot -Tpng > infra.png
inframap generate terraform.tfstate | dot -Tpng > clean.png

매 결정 기준

상황	Approach
Quick sketch	Excalidraw / Whiteboard
Reusable, version-controlled	Mermaid (md-embedded)
Strict C4 + theming	Structurizr DSL
Cloud infra reality	Terraform → Inframap
Slide deck / exec	PlantUML + C4 theme + export

기본값: Mermaid (C4 plugin) — 매 GitHub/Obsidian native rendering.

🔗 Graph

부모: Software Architecture · Documentation
변형: ADR · RFC
응용: Onboarding · Postmortem
Adjacent: Architecture Review (아키텍처 및 설계 리뷰) · Architecture_Refactor

🤖 LLM 활용

언제: text spec → Mermaid/PlantUML 의 generation, code → C4 inference. 언제 X: production diagram 의 unverified auto-generation — 매 hallucinated component 의 risk.

❌ 안티패턴

Box-and-line soup: 매 layer/audience 의 mix 의 unreadable diagram.
Stale .png in repo: 매 source-less binary — 매 update 의 X. Mermaid/DSL 의 사용.
Over-leveling: 매 Level 4 (Code) 의 manually 의 maintain — 매 IDE 의 generate.
No legend: 매 shape/color semantics 의 없는 diagram — 매 reader 의 confusion.

🧪 검증 / 중복

Verified (Simon Brown, Software Architecture for Developers; c4model.com).
신뢰도 A.

🕓 Changelog

날짜	변경
2026-05-08	Phase 1
2026-05-10	Manual cleanup — C4 model + Mermaid/Structurizr/diagrams.py patterns

5.6 KiB Raw Blame History