DongHyeonkaandClaude Opus 5
b2963105a8
The keycloak project ended with four open questions that design could not
settle. A two-VM lab was built to answer them by measurement, and this is
that material: 26 experiments, 125 raw command outputs, 22 browser captures.
Follows the import procedure in README.md.
source/ the originating repository verbatim — 78 documents, 28 SVGs,
8 manifests, plus .source-revision recording the commit
final/ the SSOT
document.md 729 lines written from the 29 experiment documents, not
concatenated: what was predicted, what was measured, and
where the measurement itself was wrong
evidence/raw 125 outputs, flattened to <experiment>__<file> because
the originals collided (01-baseline.txt appeared three
times) and the audit only globs the top level
evidence/meta one per raw file; command and exitCode are null and the
README says why rather than inventing them
evidence/browser 22 captures
assets/ three diagrams through techviz
.techviz/ their VizSpecs
A separate project rather than an addition to keycloak: the B-layer answers
that project's four questions, but the A, C and D layers are about cluster
failure, SSO and operations, and one document.md should hold one subject.
The four question records there can point here through 관계.
Recorded rather than papered over: only three of the 28 diagrams were
remade. The repository forbids hand-drawn SVG and forbids titles inside the
canvas; all 28 originals carry both, so converting them is redrawing, not
reformatting. They stay in source/ and the gap is written into the document.
verify-pipeline.py passes. audit-records.py reports no issues.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
54 lines
3.1 KiB
Markdown
54 lines
3.1 KiB
Markdown
---
|
|
kind: CONCEPT
|
|
slug: adapter-outbound-httpclient-c04
|
|
title: 뒤 규칙이 앞 규칙의 금지를 되돌리지 못한다
|
|
topic: state-machines-and-ownership
|
|
project: clean-architecture-backend-template
|
|
status: 게시 전
|
|
sourceRevision: 21234e38cdb9a926cbc92bb97a2aee2e4a7d2916
|
|
rootTreeNode: concept:adapter-outbound-httpclient-c04
|
|
evidenceCapturedOn: 2026-09-01
|
|
assets:
|
|
- key: adapter-outbound-httpclient-c04
|
|
file: ../../../final/evidence/rendered/adapter-outbound-httpclient-c04.svg
|
|
- key: adapter-outbound-httpclient-c04-diagram
|
|
file: ../../../final/assets/diagrams/adapter-outbound-httpclient-c04.svg
|
|
evidence:
|
|
- ../../../final/evidence/raw/adapter-outbound-httpclient-c04.txt
|
|
source:
|
|
- 원본 분석 절은 analysis/11-adapter-outbound-httpclient.md#L305 이다.
|
|
module: adapter-outbound-httpclient
|
|
---
|
|
|
|
# 뒤 규칙이 앞 규칙의 금지를 되돌리지 못한다
|
|
|
|
`DefaultRetryEligibilityEngine.decide`의 javadoc이 규칙이다 — 싼 절대 차단이 먼저 오고, 그다음 모호성, 그다음 상태·실패별 규칙이 온다.
|
|
|
|
## 본문
|
|
|
|
<!-- body:start -->
|
|
|
|
`DefaultRetryEligibilityEngine.decide`의 javadoc이 규칙이다 — "The order is the point. Cheap absolute blockers come first (attempts, budget, replayability, first byte, deadline, draining), then ambiguity, then status- and failure-specific rules. **A later rule can never re-enable something an earlier rule forbade.**"
|
|
|
|
## 재시도 결정의 순서
|
|
|
|
:::evidence key="adapter-outbound-httpclient-c04-diagram" alt="절대 차단 여섯과 영구 실패 범주와 증거와 상태별 규칙이 왼쪽에서 오른쪽으로 이어지는 구조" caption="재시도 결정의 순서" zoom="false"
|
|
:::
|
|
|
|
절대 차단 여섯이 먼저다 — 시도 수 소진 · 예산 소진 · 본문 재생 불가 · **첫 바이트 전달됨** · 런타임 draining · 남은 deadline이 최소 시도 예산 이하. 그다음 영구 실패 범주, 그다음 증거, 그다음 상태/실패별 규칙이다.
|
|
|
|
## DefaultRetryEligibilityEngine 참조 위치
|
|
|
|
:::evidence key="adapter-outbound-httpclient-c04" alt="코드베이스에서 DefaultRetryEligibilityEngine 를 검색한 출력 14줄. 이 기록이 세는 참조가 그 출력에 그대로 보인다." caption="DefaultRetryEligibilityEngine 코드베이스 검색 — 14줄 · exit 0" zoom="true"
|
|
:::
|
|
|
|
## HTTP 메서드가 결정에 들어 있지 않다
|
|
|
|
`RetryContext.safelyIdempotent()`가 이 모듈의 D-09를 구현한다 — HTTP 메서드는 `RetryContext`에 **아예 없다**("so a POST with a registered idempotency key and a GET against a non-idempotent RPC endpoint are both handled correctly instead of by method-name folklore"). 키 기반 멱등성은 **키가 실제로 전송됐는지**까지 요구한다. test `aKeyThatWasNeverSentDoesNotMakeARepeatSafe`가 그것을 고정한다.
|
|
|
|
## 존중한 Retry-After를 자르지 않는 이유
|
|
|
|
`Retry-After`는 남은 deadline 안에 들어갈 때만 존중된다(`allowWithin`), 그리고 존중된 `Retry-After`는 `maxBackoff`로 잘리지 **않는다** — 잘라 버리면 업스트림이 요청한 대기보다 일찍 다시 두드리게 되기 때문이다.
|
|
|
|
<!-- body:end -->
|