fix(ai): stabilize GCP Ollama alert lane

k8s/awoooi-prod/06-deployment-api.yaml

@@ -73,6 +73,12 @@ spec:
               value: "http://192.168.0.110:11436"  # 2026-05-04 ogt: GCP-B secondary via 110 nginx proxy（11436 → 34.21.145.224:11434）
             - name: OLLAMA_FALLBACK_URL
               value: "http://192.168.0.111:11434"  # 2026-05-04 ogt: 111 兜底（K8s 內網直連，GPU RTX）
+            - name: ALERT_AI_ALLOW_CLOUD_FALLBACK
+              value: "true"  # Gemini 只作 GCP-A → GCP-B → 111 全失敗後的備援
+            - name: ALERT_AI_ENFORCE_OLLAMA_FIRST
+              value: "true"  # 告警診斷強制先走 GCP-A → GCP-B → 111
+            - name: ALERT_OLLAMA_MODEL
+              value: "gemma3:4b"  # 2026-05-05 Codex: qwen3:14b 告警 JSON prompt 會拖到 504
             - name: OPENCLAW_DEFAULT_MODEL
               value: "qwen2.5:7b-instruct"
             - name: OPENCLAW_TIMEOUT