feat(wave5-p2): GovernanceAgent 4 項自檢 + Ollama 健康告警規則 + Prometheus metrics 整合

MASTER plan_complete_v3.md Wave 5 P2.2 + P2.3 完成（multiple engineers 在限額前完成代碼，補 commit）：

P2.2 — GovernanceAgent 4 項自檢:
- governance_agent.py (342 行) — 每 1 小時自檢循環:
  · trust_drift（信任度漂移檢測）
  · knowledge_degradation（知識退化檢測）
  · llm_hallucination（LLM 幻覺檢測）
  · execution_blast_radius（執行爆炸半徑檢測）
- main.py lifespan: asyncio.create_task(run_governance_loop()) 啟動
  try/except 包裹，schedule 失敗不阻斷主流程
- failover_alerter.py: alert_governance(event_type, payload) 1h dedup
  四類事件 → Telegram MarkdownV2 告警

P2.3 — Ollama 健康規則 + Prometheus Metrics:
- ops/monitoring/ollama_health_rules.yaml (148 行):
  · OllamaHealthDegraded / OllamaPrimaryDown
  · OllamaFailoverTriggered / GeminiQuotaExceeded
  · 補 Prometheus 取資料的 alert rules
- core/metrics.py (57 行):
  · GEMINI_DAILY_CALL_COUNT / GEMINI_DAILY_QUOTA Gauge
  · OLLAMA_FAILOVER_TRIGGERED_TOTAL Counter
  · OLLAMA_CURRENT_PRIMARY_IS_OLLAMA Gauge
- ollama_failover_manager.py:
  · _check_gemini_quota: 每次 check 同步更新 Gauge（讓 Prometheus 取最新值）
  · select_provider: failover 時 inc Counter + 切 Primary Gauge
  · try/except 包裹，metric 失敗不阻斷主路由

E2E 測試:
- test_failover_e2e_dispatch.py (365 行)
  完整 dispatch 路徑：health check → failover decide → alerter → metrics

Tests: 54 passed (e2e_dispatch + failover_manager + failover_alerter)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-Authored-By: Multiple Engineers (上 session Wave 5) <noreply@anthropic.com>

apps/api/src/services/ollama_failover_manager.py

@@ -424,6 +424,15 @@ class OllamaFailoverManager:
             results = await pipe.execute()
             new_count = int(results[1])           # results[1] = INCR 後新值
 
+            # 2026-04-26 P2.3 by Claude Sonnet 4.6 (tool-expert) — 刷新 Gemini Prometheus Gauge
+            # 每次 quota check 時同步更新，讓 Prometheus 取到最新值
+            try:
+                from src.core.metrics import GEMINI_DAILY_CALL_COUNT, GEMINI_DAILY_QUOTA
+                GEMINI_DAILY_CALL_COUNT.set(new_count)
+                GEMINI_DAILY_QUOTA.set(quota)
+            except Exception:
+                pass  # metric 更新失敗不阻斷主路由邏輯
+
             if new_count > quota:
                 # 已超配額（INCR 後 > quota），回退不是必要的（最多超發 1 次）
                 # 但要回傳 False 讓 router 切到 188
@@ -551,6 +560,20 @@ class OllamaFailoverManager:
             # 111 正常，無切換事件
             return
 
+        # 2026-04-26 P2.3 by Claude Sonnet 4.6 (tool-expert) — 記錄 failover Prometheus metric
+        try:
+            from src.core.metrics import (
+                OLLAMA_FAILOVER_TRIGGERED_TOTAL,
+                OLLAMA_CURRENT_PRIMARY_IS_OLLAMA,
+            )
+            OLLAMA_FAILOVER_TRIGGERED_TOTAL.labels(
+                from_provider="ollama",
+                to_provider=result.primary.provider_name,
+            ).inc()
+            OLLAMA_CURRENT_PRIMARY_IS_OLLAMA.set(0)
+        except Exception as _metric_err:
+            logger.debug("ollama_failover_metric_error", error=str(_metric_err))
+
         logger.info(
             "ollama_failover_triggered",
             service="ollama_failover",