fix(flywheel): 修補飛輪四個核心斷點，讓完整流程真正串接起來

1. incident_service.py: save_to_episodic_memory() 補寫 alertname/notification_type/alert_category
   → 之前這3欄在DB永遠NULL，LLM無alertname，Playbook匹配全失敗

2. telegram_gateway.py: Telegram批准後呼叫 execute_approved_action()
   → 之前sign_approval()只改DB狀態，380筆批准0筆真正執行kubectl指令

3. approval_execution.py: 執行成功後呼叫 resolve_incident()
   webhooks.py: auto-repair成功後呼叫 resolve_incident()
   → 之前Incident永遠停在INVESTIGATING，KM轉換永遠不觸發，Playbook=0

4. webhooks.py: TYPE-1告警短路，不進LLM
   → 之前Heartbeat/Backup/Info仍燒LLM token，產生垃圾修復建議

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

apps/api/src/services/approval_execution.py

@@ -171,6 +171,21 @@ class ApprovalExecutionService:
             except Exception as _disp_e:
                 logger.warning("disposition_record_failed", error=str(_disp_e))
 
+            # ADR-073 修補: 執行成功 → 解決 Incident → 觸發 KM 轉換
+            # 之前 RESOLVED 從未被呼叫，導致 KM 永遠不生成、Playbook 永遠是 0
+            if approval.incident_id:
+                try:
+                    from src.services.incident_service import get_incident_service
+                    _inc_svc = get_incident_service()
+                    await _inc_svc.resolve_incident(approval.incident_id)
+                    logger.info(
+                        "incident_resolved_after_execution",
+                        incident_id=approval.incident_id,
+                        approval_id=str(approval.id),
+                    )
+                except Exception as _resolve_e:
+                    logger.warning("incident_resolve_after_execution_failed", error=str(_resolve_e))
+
         else:
             logger.error(
                 "background_execution_failed",