fix(api): enforce global ollama endpoint order

apps/api/src/services/knowledge_rag_service.py

@@ -8,7 +8,7 @@ AWOOOI — Knowledge RAG Service (Phase 33, ADR-067)
   - 超過 100 筆: 執行 CREATE INDEX ivfflat (手動觸發)
 
 向量模型: bge-m3 (GCP-A/GCP-B/111 Ollama lane, 1024維)
-生成模型: qwen2.5:7b-instruct (Ollama 111)
+生成模型: qwen2.5:7b-instruct (Ollama GCP-A/GCP-B/111)
 
 leWOOOgo: Service 層只處理業務邏輯，DB 存取委派 rag_chunk_repository
 架構審查 C1 修正: 2026-04-10 Claude Sonnet 4.6 Asia/Taipei
@@ -22,7 +22,7 @@ import structlog
 
 import src.repositories.rag_chunk_repository as rag_repo
 from src.core.config import settings
-from src.services.ollama_endpoint_resolver import resolve_ollama_endpoint
+from src.services.ollama_endpoint_resolver import resolve_ollama_order
 
 logger = structlog.get_logger(__name__)
 
@@ -128,19 +128,35 @@ class KnowledgeRAGService:
     # ------------------------------------------------------------------
 
     async def _embed(self, text: str) -> list[float] | None:
-        try:
-            http = await self._get_http()
-            resp = await http.post(
-                f"{resolve_ollama_endpoint('embedding')}/api/embeddings",
-                json={
-                    "model": getattr(settings, "OLLAMA_EMBEDDING_MODEL", _EMBED_MODEL),
-                    "prompt": text,
-                },
-            )
-            if resp.status_code == 200:
-                return resp.json().get("embedding")
-        except Exception as e:
-            logger.warning("rag_embed_failed", error=str(e))
+        http = await self._get_http()
+        for endpoint in resolve_ollama_order("embedding"):
+            if not endpoint.url:
+                continue
+            try:
+                resp = await http.post(
+                    f"{endpoint.url}/api/embeddings",
+                    json={
+                        "model": getattr(settings, "OLLAMA_EMBEDDING_MODEL", _EMBED_MODEL),
+                        "prompt": text,
+                    },
+                )
+                if resp.status_code == 200:
+                    logger.debug(
+                        "rag_embed_success",
+                        provider=endpoint.provider_name,
+                    )
+                    return resp.json().get("embedding")
+                logger.warning(
+                    "rag_embed_http_error",
+                    provider=endpoint.provider_name,
+                    status=resp.status_code,
+                )
+            except Exception as e:
+                logger.warning(
+                    "rag_embed_failed",
+                    provider=endpoint.provider_name,
+                    error=str(e),
+                )
         return None
 
     async def _generate_answer(self, question: str, context: str) -> str:
@@ -150,22 +166,38 @@ class KnowledgeRAGService:
             f"=== 相關資料 ===\n{context[:6000]}\n\n"
             f"=== 問題 ===\n{question}"
         )
-        try:
-            http = await self._get_http()
-            resp = await http.post(
-                f"{resolve_ollama_endpoint('rag')}/api/generate",
-                json={
-                    "model": _GEN_MODEL,
-                    "prompt": prompt,
-                    "stream": False,
-                    "options": {"num_predict": 512, "temperature": 0.2},
-                },
-                timeout=httpx.Timeout(90.0, connect=10.0),
-            )
-            if resp.status_code == 200:
-                return resp.json().get("response", "").strip()
-        except Exception as e:
-            logger.error("rag_generate_failed", error=str(e))
+        http = await self._get_http()
+        for endpoint in resolve_ollama_order("rag"):
+            if not endpoint.url:
+                continue
+            try:
+                resp = await http.post(
+                    f"{endpoint.url}/api/generate",
+                    json={
+                        "model": _GEN_MODEL,
+                        "prompt": prompt,
+                        "stream": False,
+                        "options": {"num_predict": 512, "temperature": 0.2},
+                    },
+                    timeout=httpx.Timeout(90.0, connect=10.0),
+                )
+                if resp.status_code == 200:
+                    logger.debug(
+                        "rag_generate_success",
+                        provider=endpoint.provider_name,
+                    )
+                    return resp.json().get("response", "").strip()
+                logger.warning(
+                    "rag_generate_http_error",
+                    provider=endpoint.provider_name,
+                    status=resp.status_code,
+                )
+            except Exception as e:
+                logger.error(
+                    "rag_generate_failed",
+                    provider=endpoint.provider_name,
+                    error=str(e),
+                )
         return "⚠️ RAG 生成失敗，請稍後再試"
 
     # ------------------------------------------------------------------