feat(adr-074): M1 飛輪健康度 Exporter + M2 主機網路監控

ADR-074 M1:
  - FlywheelStatsService: 計算6項飛輪指標（Playbook數/成功率/KM向量化/alertname NULL/卡住數）
  - GET /api/v1/stats/flywheel — 六節點即時狀態（C1 前端用）
  - GET /api/v1/stats/summary — KPI 面板數據（C1 前端用）
  - GET /api/v1/stats/flywheel/metrics — Prometheus text format
  - flywheel-alerts.yaml: 5條告警規則（FlywheelPlaybookZero/ExecutionSuccessLow/KMVectorizationLow/AlertnameNullHigh/IncidentsStuck）
  - prometheus.yml: awoooi-flywheel scrape job（5分鐘間隔）

ADR-074 M2:
  - prometheus.yml: host-connectivity Blackbox TCP probe（110:22/188:22/120:6443/121:6443）
  - flywheel-alerts.yaml: HostNetworkPartition 告警規則

597 unit tests passed

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

apps/api/src/api/v1/stats.py

@@ -19,14 +19,16 @@
 # @see feedback_lewooogo_modular_enforcement.md
 # =============================================================================
 
-from typing import Annotated
+from typing import Annotated, Any
 
 from fastapi import APIRouter, Depends, Query
+from fastapi.responses import PlainTextResponse
 from pydantic import BaseModel, Field
 
 from src.services.stats_service import StatsService, get_stats_service
 from src.services.k3s_monitor_service import K3sMonitorService, get_k3s_monitor_service
 from src.services.weekly_report_service import WeeklyReportService, get_weekly_report_service
+from src.services.flywheel_stats_service import FlywheelStatsService, get_flywheel_stats_service
 
 router = APIRouter(prefix="/stats", tags=["Statistics"])
 
@@ -489,3 +491,64 @@ async def get_disposition_stats() -> DispositionResponse:
         import structlog
         structlog.get_logger(__name__).warning("disposition_stats_error", error=str(e))
         return DispositionResponse(summary=DispositionSummary())
+
+
+# =============================================================================
+# ADR-073-C C1 + ADR-074 M1 — 飛輪健康度 API
+# 2026-04-12 ogt
+# =============================================================================
+
+FlywheelStatsDep = Annotated[FlywheelStatsService, Depends(get_flywheel_stats_service)]
+
+
+@router.get(
+    "/flywheel",
+    summary="飛輪六節點即時狀態（ADR-073-C C1）",
+    response_model=None,
+)
+async def get_flywheel_stats(svc: FlywheelStatsDep) -> dict[str, Any]:
+    """
+    飛輪六節點即時狀態 + 當前流動中的告警。
+    供前端飛輪動畫元件接真實數據。
+    """
+    metrics = await svc.compute()
+    return metrics.to_flywheel_api_dict()
+
+
+@router.get(
+    "/summary",
+    summary="飛輪 KPI 摘要（ADR-073-C C1）",
+    response_model=None,
+)
+async def get_flywheel_summary(svc: FlywheelStatsDep) -> dict[str, Any]:
+    """
+    飛輪 KPI 面板數據：Playbook 數、成功率、今日處理數、KM 向量化率。
+    供前端右上角三個 KPI 卡片顯示真實數據。
+    """
+    metrics = await svc.compute()
+    return metrics.to_summary_api_dict()
+
+
+@router.get(
+    "/flywheel/metrics",
+    summary="Prometheus 飛輪健康度指標（ADR-074 M1）",
+    response_class=PlainTextResponse,
+)
+async def get_flywheel_prometheus_metrics(svc: FlywheelStatsDep) -> PlainTextResponse:
+    """
+    Prometheus text format 飛輪健康度指標。
+    Prometheus scrape target: /api/v1/stats/flywheel/metrics
+
+    Metrics:
+      awoooi_flywheel_playbook_count
+      awoooi_flywheel_execution_success_rate
+      awoooi_flywheel_km_unvectorized_count
+      awoooi_flywheel_alertname_null_rate
+      awoooi_flywheel_incidents_stuck
+      awoooi_flywheel_km_vectorized_rate
+    """
+    metrics = await svc.compute()
+    return PlainTextResponse(
+        content=metrics.to_prometheus_lines(),
+        media_type="text/plain; version=0.0.4; charset=utf-8",
+    )

apps/api/src/services/flywheel_stats_service.py

@@ -0,0 +1,386 @@
+"""
+Flywheel Stats Service — ADR-074 M1 + ADR-073-C C1
+
+飛輪健康度指標計算服務：
+  - 供 Prometheus Exporter（M1）抓取
+  - 供前端 /api/v1/stats/flywheel 即時顯示（C1）
+
+Metrics:
+  awoooi_flywheel_playbook_count          目標 ≥ 20
+  awoooi_flywheel_execution_success_rate  目標 ≥ 0.3
+  awoooi_flywheel_km_unvectorized_count   目標 = 0
+  awoooi_flywheel_alertname_null_rate     目標 = 0
+  awoooi_flywheel_incidents_stuck         目標 = 0
+
+2026-04-12 ogt (ADR-074 M1 + ADR-073-C C1)
+"""
+
+from __future__ import annotations
+
+import json
+from datetime import datetime, timedelta
+from typing import Any
+
+import structlog
+from sqlalchemy import func, select, text
+
+from src.core.redis_client import get_redis
+from src.db.base import get_db_context
+from src.db.models import IncidentRecord, KnowledgeEntryRecord
+from src.models.incident import IncidentStatus
+from src.utils.timezone import now_taipei
+
+logger = structlog.get_logger(__name__)
+
+# Redis key prefix（與 playbook_repository.py 一致）
+_PLAYBOOK_KEY_PREFIX = "playbook:"
+
+# 飛輪六節點名稱
+FLYWHEEL_NODES = [
+    "monitoring",
+    "deduplication",
+    "diagnosis",
+    "reasoning",
+    "execution",
+    "learning",
+]
+
+
+# =============================================================================
+# 核心指標資料結構
+# =============================================================================
+
+
+class FlywheelMetrics:
+    """飛輪健康度指標快照"""
+
+    def __init__(
+        self,
+        playbook_count: int,
+        execution_success_rate: float,
+        km_unvectorized_count: int,
+        alertname_null_rate: float,
+        incidents_stuck: int,
+        today_processed: int,
+        flywheel_conversions_today: int,
+        km_vectorized_rate: float,
+        node_stats: dict[str, Any],
+        current_flow: list[dict[str, Any]],
+        computed_at: datetime,
+    ) -> None:
+        self.playbook_count = playbook_count
+        self.execution_success_rate = execution_success_rate
+        self.km_unvectorized_count = km_unvectorized_count
+        self.alertname_null_rate = alertname_null_rate
+        self.incidents_stuck = incidents_stuck
+        self.today_processed = today_processed
+        self.flywheel_conversions_today = flywheel_conversions_today
+        self.km_vectorized_rate = km_vectorized_rate
+        self.node_stats = node_stats
+        self.current_flow = current_flow
+        self.computed_at = computed_at
+
+    def to_prometheus_lines(self) -> str:
+        """輸出 Prometheus text format"""
+        ts = int(self.computed_at.timestamp() * 1000)
+        lines = [
+            "# HELP awoooi_flywheel_playbook_count Total approved playbooks in Redis",
+            "# TYPE awoooi_flywheel_playbook_count gauge",
+            f"awoooi_flywheel_playbook_count {self.playbook_count} {ts}",
+            "",
+            "# HELP awoooi_flywheel_execution_success_rate Auto-repair success rate (0-1)",
+            "# TYPE awoooi_flywheel_execution_success_rate gauge",
+            f"awoooi_flywheel_execution_success_rate {self.execution_success_rate:.4f} {ts}",
+            "",
+            "# HELP awoooi_flywheel_km_unvectorized_count KM entries not yet vectorized",
+            "# TYPE awoooi_flywheel_km_unvectorized_count gauge",
+            f"awoooi_flywheel_km_unvectorized_count {self.km_unvectorized_count} {ts}",
+            "",
+            "# HELP awoooi_flywheel_alertname_null_rate Fraction of incidents with null alertname",
+            "# TYPE awoooi_flywheel_alertname_null_rate gauge",
+            f"awoooi_flywheel_alertname_null_rate {self.alertname_null_rate:.4f} {ts}",
+            "",
+            "# HELP awoooi_flywheel_incidents_stuck Incidents stuck in INVESTIGATING > 24h",
+            "# TYPE awoooi_flywheel_incidents_stuck gauge",
+            f"awoooi_flywheel_incidents_stuck {self.incidents_stuck} {ts}",
+            "",
+            "# HELP awoooi_flywheel_km_vectorized_rate Fraction of KM entries vectorized",
+            "# TYPE awoooi_flywheel_km_vectorized_rate gauge",
+            f"awoooi_flywheel_km_vectorized_rate {self.km_vectorized_rate:.4f} {ts}",
+        ]
+        return "\n".join(lines) + "\n"
+
+    def to_flywheel_api_dict(self) -> dict[str, Any]:
+        """輸出 /api/v1/stats/flywheel 格式"""
+        return {
+            "nodes": self.node_stats,
+            "current_flow": self.current_flow,
+            "computed_at": self.computed_at.isoformat(),
+        }
+
+    def to_summary_api_dict(self) -> dict[str, Any]:
+        """輸出 /api/v1/stats/summary 格式"""
+        return {
+            "playbook_count": self.playbook_count,
+            "execution_success_rate": round(self.execution_success_rate, 4),
+            "today_processed": self.today_processed,
+            "flywheel_conversions_today": self.flywheel_conversions_today,
+            "km_vectorized_rate": round(self.km_vectorized_rate, 4),
+            "km_unvectorized_count": self.km_unvectorized_count,
+            "alertname_null_rate": round(self.alertname_null_rate, 4),
+            "incidents_stuck": self.incidents_stuck,
+            "computed_at": self.computed_at.isoformat(),
+        }
+
+
+# =============================================================================
+# FlywheelStatsService
+# =============================================================================
+
+
+class FlywheelStatsService:
+    """
+    飛輪健康度指標計算服務
+
+    ADR-074 M1: 供 Prometheus /metrics/flywheel 抓取
+    ADR-073-C C1: 供前端 /api/v1/stats/flywheel 顯示
+    """
+
+    async def compute(self) -> FlywheelMetrics:
+        """計算所有飛輪指標（單次完整查詢）"""
+        now = now_taipei()
+
+        playbook_count, execution_success_rate = await self._playbook_stats()
+        (
+            km_unvectorized_count,
+            km_vectorized_rate,
+            flywheel_conversions_today,
+        ) = await self._km_stats(now)
+        (
+            alertname_null_rate,
+            incidents_stuck,
+            today_processed,
+            node_stats,
+            current_flow,
+        ) = await self._incident_stats(now)
+
+        return FlywheelMetrics(
+            playbook_count=playbook_count,
+            execution_success_rate=execution_success_rate,
+            km_unvectorized_count=km_unvectorized_count,
+            alertname_null_rate=alertname_null_rate,
+            incidents_stuck=incidents_stuck,
+            today_processed=today_processed,
+            flywheel_conversions_today=flywheel_conversions_today,
+            km_vectorized_rate=km_vectorized_rate,
+            node_stats=node_stats,
+            current_flow=current_flow,
+            computed_at=now,
+        )
+
+    # ------------------------------------------------------------------
+    # Internal helpers
+    # ------------------------------------------------------------------
+
+    async def _playbook_stats(self) -> tuple[int, float]:
+        """Playbook 數量 + 執行成功率（從 Redis）"""
+        try:
+            redis = get_redis()
+            count = 0
+            total_exec = 0
+            total_success = 0
+
+            async for key in redis.scan_iter(match=f"{_PLAYBOOK_KEY_PREFIX}PB-*", count=200):
+                raw = await redis.get(key)
+                if not raw:
+                    continue
+                try:
+                    pb = json.loads(raw)
+                    status = pb.get("status", "")
+                    if status == "approved":
+                        count += 1
+                    exec_count = pb.get("execution_count", 0) or 0
+                    success_count = pb.get("success_count", 0) or 0
+                    total_exec += exec_count
+                    total_success += success_count
+                except (json.JSONDecodeError, KeyError):
+                    continue
+
+            rate = total_success / total_exec if total_exec > 0 else 0.0
+            return count, rate
+
+        except Exception:
+            logger.exception("flywheel_stats_playbook_error")
+            return 0, 0.0
+
+    async def _km_stats(self, now: datetime) -> tuple[int, float, int]:
+        """KM 向量化率 + 今日飛輪轉化數（從 PostgreSQL）"""
+        try:
+            today_start = now.replace(hour=0, minute=0, second=0, microsecond=0)
+
+            async with get_db_context() as db:
+                # 未向量化數量
+                unvectorized_q = await db.execute(
+                    select(func.count()).where(KnowledgeEntryRecord.vectorized.is_(False))
+                )
+                unvectorized = unvectorized_q.scalar_one_or_none() or 0
+
+                # 總數
+                total_q = await db.execute(select(func.count(KnowledgeEntryRecord.id)))
+                total = total_q.scalar_one_or_none() or 0
+
+                vectorized_rate = (total - unvectorized) / total if total > 0 else 0.0
+
+                # 今日轉化數（今日建立的 KM）
+                conversions_q = await db.execute(
+                    select(func.count()).where(
+                        KnowledgeEntryRecord.created_at >= today_start
+                    )
+                )
+                conversions_today = conversions_q.scalar_one_or_none() or 0
+
+            return unvectorized, vectorized_rate, conversions_today
+
+        except Exception:
+            logger.exception("flywheel_stats_km_error")
+            return 0, 0.0, 0
+
+    async def _incident_stats(
+        self, now: datetime
+    ) -> tuple[float, int, int, dict[str, Any], list[dict[str, Any]]]:
+        """Incident 相關指標（alertname NULL 率、卡住數、今日處理數、節點狀態、當前流）"""
+        try:
+            today_start = now.replace(hour=0, minute=0, second=0, microsecond=0)
+            stuck_threshold = now - timedelta(hours=24)
+            recent_1h = now - timedelta(hours=1)
+
+            async with get_db_context() as db:
+                # alertname NULL 率
+                total_q = await db.execute(select(func.count(IncidentRecord.id)))
+                total = total_q.scalar_one_or_none() or 0
+
+                null_q = await db.execute(
+                    select(func.count()).where(IncidentRecord.alertname.is_(None))
+                )
+                null_count = null_q.scalar_one_or_none() or 0
+                alertname_null_rate = null_count / total if total > 0 else 0.0
+
+                # 卡住的 Incident（INVESTIGATING > 24h）
+                stuck_q = await db.execute(
+                    select(func.count()).where(
+                        IncidentRecord.status == IncidentStatus.INVESTIGATING.value,
+                        IncidentRecord.created_at <= stuck_threshold,
+                    )
+                )
+                incidents_stuck = stuck_q.scalar_one_or_none() or 0
+
+                # 今日處理數
+                today_q = await db.execute(
+                    select(func.count()).where(
+                        IncidentRecord.created_at >= today_start
+                    )
+                )
+                today_processed = today_q.scalar_one_or_none() or 0
+
+                # 節點狀態（監控/去重/執行）
+                recent_q = await db.execute(
+                    select(func.count()).where(
+                        IncidentRecord.created_at >= recent_1h
+                    )
+                )
+                count_1h = recent_q.scalar_one_or_none() or 0
+
+                # 自動執行成功數（今日）
+                success_q = await db.execute(
+                    text(
+                        "SELECT COUNT(*) FROM incidents WHERE created_at >= :today"
+                        " AND outcomes::text LIKE '%execution_success%true%'"
+                    ),
+                    {"today": today_start},
+                )
+                exec_success_today = success_q.scalar_one_or_none() or 0
+
+                # 當前流（最近 10 筆活躍 Incident）
+                active_q = await db.execute(
+                    select(
+                        IncidentRecord.incident_id,
+                        IncidentRecord.alertname,
+                        IncidentRecord.status,
+                        IncidentRecord.created_at,
+                    )
+                    .where(
+                        IncidentRecord.status.in_([
+                            IncidentStatus.INVESTIGATING.value,
+                            IncidentStatus.MITIGATING.value,
+                        ])
+                    )
+                    .order_by(IncidentRecord.created_at.desc())
+                    .limit(10)
+                )
+                active_rows = active_q.fetchall()
+
+            current_flow = [
+                {
+                    "incident_id": row.incident_id,
+                    "alertname": row.alertname or "unknown",
+                    "current_node": _status_to_node(row.status),
+                    "ts": row.created_at.isoformat() if row.created_at else None,
+                }
+                for row in active_rows
+            ]
+
+            node_stats = {
+                "monitoring": {
+                    "status": "active" if count_1h > 0 else "idle",
+                    "count_1h": count_1h,
+                },
+                "deduplication": {
+                    "status": "active",
+                    "dedup_window_min": 30,
+                },
+                "diagnosis": {
+                    "status": "active",
+                    "mcp_providers_used": ["k8s", "ssh", "prometheus"],
+                },
+                "reasoning": {
+                    "status": "active",
+                    "today_processed": today_processed,
+                },
+                "execution": {
+                    "status": "active",
+                    "success_today": exec_success_today,
+                },
+                "learning": {
+                    "status": "active",
+                },
+            }
+
+            return alertname_null_rate, incidents_stuck, today_processed, node_stats, current_flow
+
+        except Exception:
+            logger.exception("flywheel_stats_incident_error")
+            return 0.0, 0, 0, {n: {"status": "unknown"} for n in FLYWHEEL_NODES}, []
+
+
+def _status_to_node(status: str) -> str:
+    mapping = {
+        IncidentStatus.INVESTIGATING.value: "diagnosis",
+        IncidentStatus.MITIGATING.value: "execution",
+        IncidentStatus.RESOLVED.value: "learning",
+        IncidentStatus.CLOSED.value: "learning",
+    }
+    return mapping.get(status, "reasoning")
+
+
+# =============================================================================
+# DI 工廠
+# =============================================================================
+
+_instance: FlywheelStatsService | None = None
+
+
+def get_flywheel_stats_service() -> FlywheelStatsService:
+    global _instance
+    if _instance is None:
+        _instance = FlywheelStatsService()
+    return _instance

k8s/monitoring/flywheel-alerts.yaml

@@ -0,0 +1,105 @@
+# =============================================================================
+# 飛輪健康度告警規則 — ADR-074 M1
+# =============================================================================
+# Prometheus PrometheusRule CRD — 飛輪自監控告警
+# 數據來源：/api/v1/stats/flywheel/metrics（awoooi-flywheel scrape job）
+#
+# 部署：kubectl apply -f k8s/monitoring/flywheel-alerts.yaml
+#
+# 2026-04-12 ogt (ADR-074 M1)
+# =============================================================================
+
+apiVersion: monitoring.coreos.com/v1
+kind: PrometheusRule
+metadata:
+  name: flywheel-alerts
+  namespace: monitoring
+  labels:
+    release: prometheus
+    app: prometheus
+spec:
+  groups:
+    - name: awoooi_flywheel_health
+      interval: 5m
+      rules:
+
+        # P0: Playbook 完全沒有 → 飛輪學習節點失效
+        - alert: FlywheelPlaybookZero
+          expr: awoooi_flywheel_playbook_count == 0
+          for: 1h
+          labels:
+            severity: critical
+            alert_category: infrastructure
+            notification_type: TYPE-3
+          annotations:
+            summary: "飛輪 Playbook 數量為 0"
+            description: "Playbook 數量持續 1 小時為 0，飛輪學習節點完全失效。"
+            runbook: "執行 scripts/cold_start_playbooks.py 冷啟動"
+
+        # P0: 執行成功率極低
+        - alert: FlywheelExecutionSuccessLow
+          expr: awoooi_flywheel_execution_success_rate < 0.1
+          for: 2h
+          labels:
+            severity: warning
+            alert_category: infrastructure
+            notification_type: TYPE-3
+          annotations:
+            summary: "飛輪自動修復成功率低於 10%"
+            description: "執行成功率 {{ $value | humanizePercentage }}，低於健康基線 10%。"
+            runbook: "檢查 decision_manager 日誌，確認 target 解析和 SSH MCP 狀態"
+
+        # P0: KM 大量未向量化 → RAG 無法使用歷史案例
+        - alert: FlywheelKMVectorizationLow
+          expr: awoooi_flywheel_km_unvectorized_count > 10
+          for: 30m
+          labels:
+            severity: warning
+            alert_category: infrastructure
+            notification_type: TYPE-3
+          annotations:
+            summary: "飛輪 KM 未向量化數量 > 10"
+            description: "{{ $value }} 筆 KM 條目尚未向量化，RAG 查詢品質下降。"
+            runbook: "執行 scripts/batch_vectorize_km.py 或檢查每日 CronJob 狀態"
+
+        # P1: alertname NULL 率異常
+        - alert: FlywheelAlertnameNullHigh
+          expr: awoooi_flywheel_alertname_null_rate > 0.05
+          for: 30m
+          labels:
+            severity: warning
+            alert_category: infrastructure
+            notification_type: TYPE-3
+          annotations:
+            summary: "飛輪 alertname NULL 率超過 5%"
+            description: "alertname NULL 率 {{ $value | humanizePercentage }}，影響路由準確性。"
+            runbook: "執行 scripts/backfill_alertname.py 回填"
+
+        # P1: Incident 卡住超過 24 小時
+        - alert: FlywheelIncidentsStuck
+          expr: awoooi_flywheel_incidents_stuck > 5
+          for: 10m
+          labels:
+            severity: warning
+            alert_category: infrastructure
+            notification_type: TYPE-3
+          annotations:
+            summary: "{{ $value }} 筆 Incident 卡在 INVESTIGATING 超過 24 小時"
+            description: "大量 Incident 未推進，可能是決策引擎或 Telegram 通知阻塞。"
+
+    - name: awoooi_host_connectivity
+      interval: 60s
+      rules:
+
+        # P0: 主機間網路分區
+        - alert: HostNetworkPartition
+          expr: probe_success{job="host-connectivity"} == 0
+          for: 5m
+          labels:
+            severity: critical
+            alert_category: infrastructure
+            notification_type: TYPE-3
+          annotations:
+            summary: "主機 {{ $labels.instance }} 無法連通"
+            description: "TCP probe 到 {{ $labels.instance }} 失敗超過 5 分鐘，可能發生網路分區。"
+            runbook: "SSH 檢查路由和防火牆規則"

k8s/monitoring/prometheus.yml

@@ -156,6 +156,37 @@ scrape_configs:
           service: 'awoooi-api'
           env: 'prod'
 
+  # === ADR-074 M1: 飛輪健康度指標 (2026-04-12 ogt) ===
+  - job_name: 'awoooi-flywheel'
+    scrape_interval: 5m
+    metrics_path: /api/v1/stats/flywheel/metrics
+    static_configs:
+      - targets: ['192.168.0.125:32334']
+        labels:
+          host: '125'
+          service: 'awoooi-flywheel'
+          env: 'prod'
+
+  # === ADR-074 M2: 主機間網路連通性 (2026-04-12 ogt) ===
+  - job_name: 'host-connectivity'
+    scrape_interval: 60s
+    metrics_path: /probe
+    params:
+      module: [tcp_connect]
+    static_configs:
+      - targets:
+          - 192.168.0.110:22
+          - 192.168.0.188:22
+          - 192.168.0.120:6443
+          - 192.168.0.121:6443
+    relabel_configs:
+      - source_labels: [__address__]
+        target_label: __param_target
+      - source_labels: [__param_target]
+        target_label: instance
+      - target_label: __address__
+        replacement: 192.168.0.188:9115
+
   # === Sprint 5.2 Plan B: PostgreSQL Exporter (2026-04-08 Claude Sonnet 4.6) ===
   - job_name: 'postgres-exporter'
     scrape_interval: 30s