Files

T

koriweb d8a80f6272 chore(wiki): dangling 링크 canonical 정규화 (768파일/1200건)

이름만 다른(표기 변형) [[위키링크]]를 대상 문서의 canonical 제목으로 치환해
끊겼던 1,200개 링크를 연결. 제목/파일명 정규화 일치만 적용하고 별칭 매칭은
과병합 위험으로 제외(애매성 가드). 원본은 _link_reconcile_backup/ 에 백업.
도구: Datacollect/scripts/link_reconcile_apply.mjs

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

2026-06-08 12:24:15 +09:00

3.8 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

Flame Chart

매 한 줄

"매 시간축 stack frame 의 visual profile — width=duration, depth=call stack". Brendan Gregg (2013) 의 flame graph 매 origin. 2026 시점 Chrome DevTools Performance panel · Node --prof · Linux perf · Speedscope 매 standard tooling.

매 핵심

매 Flame Chart vs Flame Graph

Flame Chart: x축 = 매 wall-clock time (left→right chronological). DevTools default.
Flame Graph: x축 = 매 aggregated time (alphabetical), 매 hot path 발견.

매 reading rules

매 wide bar = 매 long-running function.
매 tall stack = 매 deep call chain (recursion · over-abstraction signal).
매 plateau = 매 single function dominate.

매 응용

JS main-thread bottleneck 식별.
React render 의 expensive component 추적.
Backend RPC handler latency 분해.

💻 패턴

Chrome DevTools 캡처

// 1. DevTools → Performance → Record
// 2. Reproduce slow interaction
// 3. Stop → bottom-up / call tree / flame chart 탭

// Programmatic mark
performance.mark("render-start")
expensiveRender()
performance.mark("render-end")
performance.measure("render", "render-start", "render-end")

Node.js CPU profile

node --cpu-prof --cpu-prof-dir=./profiles app.js
# .cpuprofile 의 Chrome DevTools "Performance" 탭에 drag-drop

0x flame graph 생성

npx 0x -o server.js
# HTML 매 자동 open — interactive flame graph

React Profiler API

import { Profiler } from "react"

function onRender(id, phase, actualDuration) {
  if (actualDuration > 16) console.warn(`slow ${id}: ${actualDuration}ms`)
}

<Profiler id="UserList" onRender={onRender}>
  <UserList />
</Profiler>

Speedscope (universal viewer)

# Node, Chrome, py-spy, perf 매 모두 import
npx speedscope profile.cpuprofile

매 결정 기준

상황	Approach
Browser frontend lag	Chrome DevTools Performance
Node server CPU spike	`--cpu-prof` + DevTools
Production hot-path discovery	Flame graph (aggregated)
Specific user interaction	Flame chart (chronological)

기본값: Chrome DevTools Performance panel — 매 60% case cover.

🔗 Graph

부모: Performance_Profiling_and_Memory · Chrome DevTools(크롬 개발자 도구)
변형: 할당 타임라인(Allocation Timeline)
응용: Nodejs 성능 디버깅 · SPA 라우트 전환 성능 최적화
Adjacent: Core Web Vitals Optimization (INP, LCP, CLS) · Google Lighthouse

🤖 LLM 활용

언제: 매 latency / jank / unknown bottleneck 발견 → flame chart 의 first tool. 언제 X: 매 memory leak (heap snapshot 사용) · 매 network waterfall (Network panel).

❌ 안티패턴

Optimize the wide-but-shallow bar: 매 actually unimportant 일 수 있음 — call tree로 cross-check.
Sampling rate 신뢰: 매 short bursts (<1ms) 의 missed — performance.mark precise.
Production 에서 always-on profiling: 매 overhead — sampling profiler (e.g. pprof) 만 OK.

🧪 검증 / 중복

Verified (Brendan Gregg flame graph paper · Chrome DevTools docs 2026).
신뢰도 A.

🕓 Changelog

날짜	변경
2026-05-08	Phase 1
2026-05-10	Manual cleanup — DevTools/0x/Speedscope tooling 매 설명

3.8 KiB Raw Blame History