feat(web): surface callback trace backlog work item

2026-05-25 22:00:42 +08:00
parent 2c058e5adf
commit 44f48b68fe
4 changed files with 149 additions and 0 deletions
--- a/apps/web/messages/en.json
+++ b/apps/web/messages/en.json
@@ -2225,6 +2225,7 @@
        "remediationQueue": "Every degraded / failed / timeout row must map to replay, reverify, ticket, or manual review",
        "telegramCallbacks": "Detail and history buttons cannot depend only on Redis TTL or stale snapshots",
        "callbackOwnerReview": "Telegram detail/history callbacks without a KM owner-review link must become trackable work items",
+        "callbackTraceRecoveryBacklog": "Callback trace gaps must show recovery signal, 24h decay, and the backlog next step",
        "ciSecretHygiene": "Workflows must not mount 機密設定 in step env / action inputs; historical logs still need rotation and retention governance",
        "governanceDispatch": "Governance alerts must enter dispatch and expose skipped / pending / repaired",
        "knowledgeHealthcheck": "knowledge_degradation must show Hermes / OpenClaw / ElephantAlpha ownership, current stage, and owner review point",
@@ -2270,6 +2271,10 @@
        "callbackOwnerReviewAutomation": "Automation: {state}; safe auto-repair={safe}",
        "callbackOwnerReviewBlocker": "Blocker: {reason}",
        "callbackOwnerReviewEmpty": "Recent callback evidence is matched or no data is available yet",
+        "callbackTraceRecoveryBacklog": "Callback trace backlog: missing trace {missing}; 1h {recent1h}; 24h {recent24h}; traced after gap {recovered}; recovery {status}",
+        "callbackTraceRecoveryDecision": "Decision: {gap}; next: {next}",
+        "callbackTraceRecoveryLatest": "Last gap: {missing}; recovery first: {first}; recovery latest: {latest}",
+        "callbackTraceRecoveryUnavailable": "Callback trace recovery summary has not returned yet",
        "ciSecretHygiene": "Repo-controlled step env / action input exposure is guarded; key rotation and log retention remain",
        "governance": "Unresolved governance alerts: {unresolved}; 待派送: {queued}",
        "governanceUnavailable": "Governance events API is not responding; 待派送: {queued}",
--- a/apps/web/messages/zh-TW.json
+++ b/apps/web/messages/zh-TW.json
@@ -2226,6 +2226,7 @@
        "remediationQueue": "每筆 degraded / failed / timeout 都必須映射到重跑、重驗、Ticket 或人工檢查",
        "telegramCallbacks": "按下詳情與歷史不能再只依賴 Redis TTL 或舊快照",
        "callbackOwnerReview": "Telegram 詳情 / 歷史若未連到 KM owner-review，必須變成可追蹤工作項",
+        "callbackTraceRecoveryBacklog": "Callback trace 缺口必須顯示復原訊號、24h decay 與 backlog 下一步",
        "ciSecretHygiene": "workflow 不可再把 機密設定 掛在 step env / action input；歷史 log 需另做輪換與保留期治理",
        "governanceDispatch": "治理告警必須進 dispatch，並標示 skipped / pending / repaired",
        "knowledgeHealthcheck": "knowledge_degradation 必須顯示 Hermes / OpenClaw / ElephantAlpha 分工、目前階段與 owner 審核點",
@@ -2271,6 +2272,10 @@
        "callbackOwnerReviewAutomation": "自動化：{state}；可安全自動修復={safe}",
        "callbackOwnerReviewBlocker": "卡點：{reason}",
        "callbackOwnerReviewEmpty": "近期 callback evidence 均已匹配或尚無資料",
+        "callbackTraceRecoveryBacklog": "Callback trace backlog：缺 trace {missing}；1h {recent1h}；24h {recent24h}；gap 後 traced {recovered}；復原 {status}",
+        "callbackTraceRecoveryDecision": "判讀：{gap}；下一步：{next}",
+        "callbackTraceRecoveryLatest": "最後缺口：{missing}；復原首筆：{first}；復原最新：{latest}",
+        "callbackTraceRecoveryUnavailable": "Callback trace recovery summary 尚未回傳",
        "ciSecretHygiene": "repo 可控 step env / action input 泄漏面已加 guard；仍需 key rotation 與 log retention 收斂",
        "governance": "未解治理告警：{unresolved}；待派送：{queued}",
        "governanceUnavailable": "治理事件 API 目前無法回應；待派送：{queued}",
--- a/apps/web/src/app/[locale]/awooop/work-items/page.tsx
+++ b/apps/web/src/app/[locale]/awooop/work-items/page.tsx
@@ -827,11 +827,25 @@ type CallbackReplyWorkItemEvent = {
  km_stale_completion_summary?: KmStaleCallbackCompletionSummary | null;
 };

+type CallbackReplyAuditSummary = {
+  outbound_reply_markup_missing_trace_ref_total?: number;
+  outbound_reply_markup_missing_trace_ref_recent_1h_total?: number;
+  outbound_reply_markup_missing_trace_ref_recent_24h_total?: number;
+  outbound_reply_markup_missing_trace_ref_latest_sent_at?: string | null;
+  outbound_reply_markup_trace_ref_gap_status?: string | null;
+  outbound_reply_markup_trace_ref_gap_next_action?: string | null;
+  outbound_reply_markup_trace_ref_after_gap_total?: number;
+  outbound_reply_markup_trace_ref_after_gap_first_sent_at?: string | null;
+  outbound_reply_markup_trace_ref_after_gap_latest_sent_at?: string | null;
+  outbound_reply_markup_trace_ref_gap_recovery_status?: string | null;
+};
+
 type CallbackRepliesWorkItemResponse = {
  items?: CallbackReplyWorkItemEvent[];
  total: number;
  page: number;
  per_page: number;
+  summary?: CallbackReplyAuditSummary | null;
 };

 type AiRouteRepairEvidence = {
@@ -1610,6 +1624,27 @@ function callbackOwnerReviewOpenEvents(
  });
 }

+function callbackTraceRecoveryStatus(
+  summary: CallbackReplyAuditSummary | null | undefined
+): WorkStatus {
+  if (!summary) {
+    return "blocked";
+  }
+  const missingTrace = summary.outbound_reply_markup_missing_trace_ref_total ?? 0;
+  if (missingTrace <= 0) {
+    return "live";
+  }
+  const gapStatus = summary.outbound_reply_markup_trace_ref_gap_status;
+  const recoveryStatus = summary.outbound_reply_markup_trace_ref_gap_recovery_status;
+  if (gapStatus === "active_gap" || recoveryStatus === "no_recovery_signal") {
+    return "blocked";
+  }
+  if (recoveryStatus === "recovered_after_gap") {
+    return "in_progress";
+  }
+  return "watching";
+}
+
 function buildWorkItems(
  telemetry: Telemetry,
  t: ReturnType<typeof useTranslations>
@@ -1643,6 +1678,7 @@ function buildWorkItems(
    latestCallbackOwnerReview?.km_stale_completion_summary ?? null;
  const latestCallbackWorkItem = latestCallbackSummary?.work_item ?? null;
  const latestCallbackTriage = latestCallbackWorkItem?.triage ?? null;
+  const callbackTraceSummary = telemetry.callbackReplies?.summary ?? null;
  const aiRoute = telemetry.aiRouteStatus;
  const aiRouteRepairEvidence = aiRoute?.repair_evidence ?? null;
  const aiRouteWorkItem = aiRouteRepairEvidence?.work_item ?? null;
@@ -1993,6 +2029,58 @@ function buildWorkItems(
        : [t("evidence.callbackOwnerReviewEmpty")],
      href: latestCallbackWorkItem?.target_href ?? "/awooop/runs",
    },
+    {
+      id: "callbackTraceRecoveryBacklog",
+      phase: "T195",
+      status: callbackTraceRecoveryStatus(callbackTraceSummary),
+      surfaceKey: "workItems",
+      source: "/api/v1/platform/runs/callback-replies summary",
+      gateKey: "callbackTraceRecoveryBacklog",
+      evidence: t("evidence.callbackTraceRecoveryBacklog", {
+        missing:
+          callbackTraceSummary?.outbound_reply_markup_missing_trace_ref_total ??
+          0,
+        recent1h:
+          callbackTraceSummary
+            ?.outbound_reply_markup_missing_trace_ref_recent_1h_total ?? 0,
+        recent24h:
+          callbackTraceSummary
+            ?.outbound_reply_markup_missing_trace_ref_recent_24h_total ?? 0,
+        recovered:
+          callbackTraceSummary?.outbound_reply_markup_trace_ref_after_gap_total ??
+          0,
+        status:
+          callbackTraceSummary
+            ?.outbound_reply_markup_trace_ref_gap_recovery_status ?? "--",
+      }),
+      evidenceDetails: callbackTraceSummary
+        ? [
+            t("evidence.callbackTraceRecoveryDecision", {
+              gap:
+                callbackTraceSummary.outbound_reply_markup_trace_ref_gap_status ??
+                "--",
+              next:
+                callbackTraceSummary
+                  .outbound_reply_markup_trace_ref_gap_next_action ?? "--",
+            }),
+            t("evidence.callbackTraceRecoveryLatest", {
+              missing:
+                callbackTraceSummary
+                  .outbound_reply_markup_missing_trace_ref_latest_sent_at ??
+                "--",
+              first:
+                callbackTraceSummary
+                  .outbound_reply_markup_trace_ref_after_gap_first_sent_at ??
+                "--",
+              latest:
+                callbackTraceSummary
+                  .outbound_reply_markup_trace_ref_after_gap_latest_sent_at ??
+                "--",
+            }),
+          ]
+        : [t("evidence.callbackTraceRecoveryUnavailable")],
+      href: "/awooop/runs?project_id=awoooi",
+    },
    {
      id: "ciSecretHygiene",
      phase: "T44",
--- a/docs/LOGBOOK.md
+++ b/docs/LOGBOOK.md
@@ -21620,3 +21620,54 @@ GET /api/v1/health
 - KM governance：約 84.6%。
 - AI Provider lane visibility：約 92.2%。
 - 完整 AI 自動化管理產品化：約 98.15%。
+
+### 2026-05-25 — T195 Callback trace recovery backlog work item（pre-deploy）
+
+**背景**：T194 已在 Runs / TG Callback Evidence 顯示 `recovered_after_gap`，但它
+仍只是摘要文字。T195 把同一份 callback summary 接進 AwoooP Work Items，讓
+`recovered_after_gap + 24h decay` 進入 backlog 視角，而不是靠值班者記住。
+
+**完成變更**：
+
+- AwoooP Work Items 讀取 `/api/v1/platform/runs/callback-replies` 的 `summary`。
+- 新增 `callbackTraceRecoveryBacklog` 工作項投影：
+  - `missing_trace_ref_total`
+  - `recent_1h / recent_24h`
+  - `trace_ref_after_gap_total`
+  - `trace_ref_gap_recovery_status`
+  - `trace_ref_gap_status / next_action`
+- 狀態規則：
+  - 無 summary：`blocked`。
+  - 缺 trace = 0：`live`。
+  - `active_gap` 或 `no_recovery_signal`：`blocked`。
+  - `recovered_after_gap`：`in_progress`，代表舊 backlog 正在 24h decay。
+  - 其他：`watching`。
+
+**local validation（完成）**：
+
+```text
+jq empty apps/web/messages/zh-TW.json apps/web/messages/en.json
+pnpm --dir apps/web exec tsc --noEmit --tsBuildInfoFile /tmp/awoooi-t195-tsconfig.tsbuildinfo
+  pass
+pnpm --dir apps/web lint -- --file 'src/app/[locale]/awooop/work-items/page.tsx'
+  pass; no warnings
+NEXT_PUBLIC_API_URL=https://awoooi.wooo.work pnpm --dir apps/web run build
+  pass; Sentry global-error / instrumentation-client warnings are pre-existing
+git diff --check
+  pass
+```
+
+**目前整體進度（pre-deploy）**：
+
+- AwoooP 告警可觀測鏈：約 99.88%。
+- 低風險自動修復閉環：約 95.9%。
+- 前端 AI 自動化管理介面同步：約 99.8%。
+- Telegram outbound / callback DB coverage 可視化：約 99.88%。
+- callback / DB replayability：約 99.1%。
+- Work Items / backlog 可追蹤性：約 96.5%。
+- MCP / 自建 MCP 可視化：約 95.1%。
+- Sentry / SigNoz source correlation：約 94.5%。
+- Ansible / PlayBook 可視化：約 92.6%。
+- KM governance：約 84.6%。
+- AI Provider lane visibility：約 92.2%。
+- 完整 AI 自動化管理產品化：約 98.2%。