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>
4.0 KiB
kind, slug, title, topic, project, status, sourceRevision, rootTreeNode, evidenceCapturedOn, assets, evidence, source, module
| kind | slug | title | topic | project | status | sourceRevision | rootTreeNode | evidenceCapturedOn | assets | evidence | source | module | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CONCEPT | messaging-reliability-api-c03 | 이 enum의 의미가 저장소 규칙 하나의 존재 이유다 | state-machines-and-ownership | clean-architecture-backend-template | 게시 전 | 21234e38cdb9a926cbc92bb97a2aee2e4a7d2916 | concept:messaging-reliability-api-c03 | 2026-09-01 |
|
|
|
messaging-reliability-api |
이 enum의 의미가 저장소 규칙 하나의 존재 이유다
OutboxLease의 fencing token이 이 leaf에서 가장 중요한 안전 장치이고, OutboxStatus.FAILED의 의미가 애플리케이션 쪽 동명 enum과 반대다.
관계
- 계약만 담는 leaf도 계약의 거절 조건은 자기 레인에서 검증한다 같은 분석 리프에서 끌어낸 규칙이다.
- record의
equals를 좁히면 이유를 적는다 같은 분석 리프에서 끌어낸 규칙이다. - 호출 컨텍스트가 계약이면 그 컨텍스트를 검증할 수단을 함께 정한다 같은 분석 리프에서 끌어낸 규칙이다.
본문
OutboxLease가 이 leaf에서 가장 중요한 안전 장치이고, 이전 결함이 javadoc에 통째로 있다.
"The port used to take a
MessageIdfor every terminal transition, so a write said which row to change and nothing about which claim it belonged to. A relay that stalled past its lease could still recordAMBIGUOUSover thePUBLISHEDanother relay had already written, and the row became claimable again — one message, published twice, by a system whose whole purpose is to publish it once. The token is the part that makes staleness detectable. It increases on every claim, so a superseded relay holds a number the row no longer has and its update matches zero rows."
token < 1을 거절하는 이유도 적혀 있다 — "a claim's token starts at 1; 0 is the value of a row nobody has claimed". expiredAt(now)가 !now.isBefore(expiresAt)다.
청구된 행이 갈리는 네 종착
:::evidence key="messaging-reliability-api-c03-diagram" alt="IN_FLIGHT 에서 PUBLISHED 와 AMBIGUOUS 와 FAILED 와 EXHAUSTED 네 상자로 화살표가 나가는 구조" caption="청구된 행이 갈리는 네 종착" zoom="false" :::
PENDING → IN_FLIGHT → PUBLISHED / AMBIGUOUS / FAILED / EXHAUSTED. 두 쌍의 구분이 각각 이유를 갖는다. OutboxRepository.markExhausted의 javadoc이 같은 말을 반복한다 — "The first needs a fix, the second a redrive."
OutboxRepository 참조 위치
:::evidence key="messaging-reliability-api-c03" alt="코드베이스에서 OutboxRepository 를 검색한 출력 35줄. 이 기록이 세는 참조가 그 출력에 그대로 보인다." caption="OutboxRepository 코드베이스 검색 — 35줄 · exit 0" zoom="true" :::
이름이 같고 의미가 반대인 enum
FAILED의 의미가 애플리케이션 쪽 동명 enum과 반대다. CleanArchitectureTest.APPLICATION_DOES_NOT_DEPEND_ON_THE_MESSAGING_PLATFORM의 .because(...)가 그것을 ArchUnit 규칙의 근거로 든다 — "its OutboxStatus.FAILED means the opposite of the legacy OutboxEventStatus.FAILED, so the two models cannot be mixed by name without inverting retryable and terminal." 즉 이 enum의 의미가 저장소 규칙 하나의 존재 이유다.
정산하면 안 되는 경우를 상수가 들고 있다
safeToSettle 플래그가 상수에 붙어 있다. 세 번째의 javadoc이 결론을 적는다 — "Do not settle. The other transaction may still roll back, and this delivery is the only remaining copy."