docs: give the twelve experiments that had no architecture diagram one

An audit against the standard the series set — concepts, procedure,
commands, architecture diagram, evidence table, terminal output — found the
three new experiments met it while twelve of the original ones had no
diagram at all: A-0, A-1, A-3, A-4, A-5, A-6, A-8, B-0, B-2, B-7, C-2, D-2.

Each now has one drawn from what that experiment actually found, not filler:
A-0 shows sharing going through PostgreSQL rather than between the caches;
A-3 the gap between the 200 and the WAL flush, with both failed injections;
A-5 the three silent injection failures; A-6 the two places latency is
multiplied; B-0 the repository keyed by principal with no session id; B-2
the primary key that causes the overwrite; D-2 why the rolling update
stopped the accident halfway.

Also corrected the index's stale claim of 11 experiments without a
screenshot — it is 14, and the reason is recorded: those experiments were
measured from terminals, the database and logs, and the observability stack
does not scrape Redis, the BFF or PostgreSQL, so there is no console to
photograph.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

docs/diagrams/a5-asymmetric-partition.svg

@@ -0,0 +1,49 @@
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 860 420" font-family="-apple-system,BlinkMacSystemFont,Segoe UI,Helvetica,sans-serif">
+<style>
+.box{fill:#fff;stroke:#8c959f;stroke-width:1.4;rx:6}
+.dead{fill:#fff5f5;stroke:#cf222e;stroke-width:1.6;stroke-dasharray:5 3;rx:6}
+.ok{fill:#f6fdf6;stroke:#1a7f37;stroke-width:1.6;rx:6}
+.warn{fill:#fff8f0;stroke:#bf8700;stroke-width:1.5;rx:6}
+.t{font-size:12px;fill:#1f2328}
+.tb{font-size:12.5px;fill:#1f2328;font-weight:600}
+.s{font-size:10.5px;fill:#656d76}
+.ttl{font-size:14px;fill:#1f2328;font-weight:600}
+.r{stroke:#cf222e;stroke-width:1.6;fill:none}
+.g{stroke:#1a7f37;stroke-width:1.6;fill:none}
+.k{stroke:#8c959f;stroke-width:1.3;fill:none}
+</style>
+<defs>
+<marker id="a" markerWidth="8" markerHeight="8" refX="7" refY="3" orient="auto"><path d="M0,0 L7,3 L0,6 z" fill="#8c959f"/></marker>
+<marker id="ar" markerWidth="8" markerHeight="8" refX="7" refY="3" orient="auto"><path d="M0,0 L7,3 L0,6 z" fill="#cf222e"/></marker>
+<marker id="ag" markerWidth="8" markerHeight="8" refX="7" refY="3" orient="auto"><path d="M0,0 L7,3 L0,6 z" fill="#1a7f37"/></marker>
+</defs>
+<text x="20" y="26" class="ttl">A-5 — 비대칭 차단은 클러스터를 가르지 못한다. 그리고 갈라도 서비스는 계속된다</text>
+<text x="20" y="44" class="s">주입을 세 번 실패했고, 세 번 모두 「아무 일도 없었다」로 보였다</text>
+<rect x="30" y="72" width="390" height="112" class="box"/>
+<text x="46" y="94" class="tb">① 한 방향만 차단</text>
+<rect x="46" y="106" width="130" height="34" class="box"/>
+<text x="58" y="128" class="t">keycloak-0</text>
+<rect x="290" y="106" width="116" height="34" class="box"/>
+<text x="302" y="128" class="t">keycloak-1</text>
+<path d="M178 116 L286 116" class="r" stroke-dasharray="4 3" marker-end="url(#ar)"/>
+<path d="M286 132 L178 132" class="g" marker-end="url(#ag)"/>
+<text x="46" y="164" class="s" fill="#1a7f37">열린 방향으로 재연결한다 → cluster_size 2 유지. 가르지 못한다</text>
+<rect x="440" y="72" width="390" height="112" class="box"/>
+<text x="456" y="94" class="tb">② 양방향 완전 차단</text>
+<rect x="456" y="106" width="130" height="34" class="dead"/>
+<text x="468" y="128" class="t">keycloak-0</text>
+<rect x="700" y="106" width="116" height="34" class="ok"/>
+<text x="712" y="128" class="t">keycloak-1</text>
+<path d="M588 116 L696 116" class="r" stroke-dasharray="4 3" marker-end="url(#ar)"/>
+<path d="M696 132 L588 132" class="r" stroke-dasharray="4 3" marker-end="url(#ar)"/>
+<text x="456" y="164" class="s">양쪽 모두 멤버 1개. 그런데 <tspan fill="#cf222e" font-weight="600">한쪽만 DOWN</tspan> 이 된다 — 코디네이터 쪽이 살아남는다</text>
+<rect x="30" y="204" width="800" height="48" class="ok"/>
+<text x="46" y="226" class="tb">그래서 전면 장애 경로가 없다 — 외부 응답 200 유지</text>
+<text x="46" y="244" class="s">분단된 쪽이 스스로 readiness 를 떨어뜨려 로드밸런서에서 빠지고, 남은 쪽이 계속 응답한다</text>
+<rect x="30" y="272" width="800" height="130" class="warn"/>
+<text x="46" y="294" class="tb">★ 세 번의 주입 실패 — 전부 「조용히」 실패했다</text>
+<text x="46" y="318" class="s">1. iptables -I FORWARD 1 이 무시됐다 — kube-router 가 자기 체인을 FORWARD 맨 위에 다시 끼워 넣는다 (패킷 0)</text>
+<text x="46" y="338" class="s">2. 방향이 뒤집혀 있었다 — JGroups 의 client/server 역할은 재시작마다 바뀐다. raw 규칙이 엉뚱한 노드에 걸렸다</text>
+<text x="46" y="358" class="s">3. dev eth0 이 없다 — Debian 은 enp1s0 이고, 게다가 flannel VXLAN 이 이미 캡슐화해 파드 IP 가 안 보인다</text>
+<text x="46" y="382" class="s" fill="#cf222e">해결: raw 테이블 PREROUTING(conntrack 보다 먼저) + flannel.1 인터페이스에서 필터</text>
+</svg>