diff --git a/docs/LOGBOOK.md b/docs/LOGBOOK.md index 9d54e1a3..5f17cdd4 100644 --- a/docs/LOGBOOK.md +++ b/docs/LOGBOOK.md @@ -22112,6 +22112,94 @@ production frontend smoke: - 完整 AI 自動化管理產品化:約 98.7%,但「真正 AI 自動修復閉環」仍需補齊 execution / repair / approval / learning evidence 與 Ansible runtime gate。 +### 2026-05-26 — T202 Homepage blueprint live evidence binding(post-deploy) + +**背景**:T201 讓首頁 AI 自動化作戰圖可 deep-link 到每個流程節點,但 Stage Inspector +仍偏靜態說明,無法直接回答 Telegram / Sentry / SigNoz / MCP / KM 等證據目前實際落在哪裡。 +本階段把首頁節點改成讀既有 production read-only API,不新增後端寫入、不造假數據。 + +**完成變更**: + +- 首頁 Live Blueprint 每個流程節點新增 `Live Evidence`: + - Signal:`/api/v1/platform/events/dossier/coverage`,顯示 source refs / missing refs / Alert / Sentry / SigNoz。 + - Intake:`/api/v1/platform/runs/list` + `/api/v1/platform/cicd/events`,顯示 runs / linked / 最新 CI/CD。 + - AI Route:`/api/v1/platform/ai-route-status`,顯示 lane / provider / skipped lanes / operator action。 + - MCP:`/api/v1/platform/runs/list` + status-chain 摘要,顯示 MCP observations / success / failed / server / route。 + - PlayBook:truth-chain quality + dossier recurrence,顯示 gate / automation gaps / work items。 + - Ansible:truth-chain quality,顯示 check-mode / pending / blocker / candidates / operations。 + - Approval:approval + truth-chain 摘要,顯示 pending / verified / human gates / auto-repair records。 + - Verify / KM:`/api/v1/ai/governance/km-stale-owner-review-burndown`,顯示 stale ratio / owner review / 距離門檻。 +- Stage Inspector 新增 live evidence 欄位:指標、細節、讀取來源、更新時間。 +- 首頁 fetch 改成 per-source independent loading: + - 不再用單一 `Promise.all` 卡住整個 blueprint。 + - 任一 API 先回來就先更新自己的節點,慢來源只影響自己的欄位。 +- 新增 zh-TW / en i18n,新增產品文案未硬寫在 JSX。 + +**validation / deployment evidence**: + +```text +local: + jq empty apps/web/messages/zh-TW.json apps/web/messages/en.json + pnpm --dir apps/web exec tsc --noEmit --tsBuildInfoFile /tmp/awoooi-home-live-evidence-tsconfig.tsbuildinfo + pnpm --dir apps/web exec tsc --noEmit --tsBuildInfoFile /tmp/awoooi-home-live-evidence-independent-tsconfig.tsbuildinfo + pnpm --dir apps/web lint -- --file 'src/app/[locale]/page.tsx' + pass; remaining warnings are pre-existing any / literal warnings in the old page + git diff --check + NEXT_PUBLIC_API_URL=https://awoooi.wooo.work pnpm --dir apps/web run build + pass; /[locale] = 26.6 kB, Sentry warnings are pre-existing + Playwright localhost structural smoke: + blueprint_stage=signal/mcp/verify + Live Evidence visible = true + source route visible = true + document/body horizontal overflow = false + content scrolling works + +gitea: + 2190 CD success + 2191 code-review success + 2192 CD success + 2193 code-review success + +commits: + 320718aa feat(web): bind homepage blueprint to live evidence + 01c6cb29 fix(web): stream homepage evidence sources independently + 5cfee5cf chore(cd): deploy 320718a [skip ci] + b39fded8 chore(cd): deploy 01c6cb2 [skip ci] + +production frontend smoke: + https://awoooi.wooo.work/zh-TW?blueprint_stage=signal#blueprint-stage-inspector + desktop readyMs=8225, source_count=100, refs=500, missing=0, duplicates=46, Alert=200, Sentry=0, SigNoz=0 + mobile readyMs=24748, document/body scrollWidth=390, viewport=390, inner scroll works + https://awoooi.wooo.work/zh-TW?blueprint_stage=mcp#blueprint-stage-inspector + desktop readyMs=41250, MCP observations=42, success=12, failed=30, server=ssh_host + mobile readyMs=34222, document/body scrollWidth=390, viewport=390, inner scroll works + https://awoooi.wooo.work/zh-TW?blueprint_stage=verify#blueprint-stage-inspector + desktop readyMs=2783, stale=1567, stale_ratio=51.1%, pending_owner_reviews=10, completed=1, remaining=954 + mobile readyMs=9835, document/body scrollWidth=390, viewport=390, inner scroll works + all smoke: + console_errors = 0 + unexpected failed requests = 0 + Live Evidence has no 讀取中 / 未回應 after ready + desktop document/body scrollWidth=1440, viewport=1440 +``` + +**目前整體進度(post-deploy)**: + +- 首頁產品化工作地圖:約 99.2%。 +- 首頁可用性 / 導航 / 滾動:約 99.8%。 +- 專業圖像化呈現:約 96.5%。 +- 前端 AI 自動化管理介面同步:約 99.95%。 +- Telegram outbound / callback DB coverage 可視化:約 99.9%。 +- Work Items / backlog 可追蹤性:約 98.0%。 +- AI Provider lane visibility:約 94.0%。 +- MCP / 自建 MCP 可視化:約 96.5%。 +- Sentry / SigNoz source correlation:約 95.0%(目前 coverage 顯示 Sentry/SigNoz refs=0,代表仍需讓告警來源補齊引用)。 +- KM governance:約 85.2%(stale ratio 仍約 51.1%,Hermes draft + owner review 需要繼續消化)。 +- Ansible / PlayBook 自動執行:約 0% runtime-ready(仍受 `ansible_playbook_binary_missing` gate 阻塞)。 +- 24h 完整自動修復 production claim:0%(不能宣稱真正全自動修復閉環)。 +- 完整 AI 自動化管理產品化:約 99.1%,但「真正 AI 自動修復閉環」仍需補齊 + execution / repair / approval / learning evidence 與 Ansible runtime gate。 + ### 2026-05-26 — T201 Homepage blueprint drill-down and mobile evidence wrapping(post-deploy) **背景**:T200 已把首頁從空洞入口升級成可見的 AI 自動化作戰圖,但仍只能「看流程」,