ok

app/api/v1/chat.py

@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+import logging
+import time
 from typing import Any, Dict, List, Optional
 
 from fastapi import APIRouter, Depends, HTTPException, Request
@@ -7,6 +9,8 @@ from fastapi import APIRouter, Depends, HTTPException, Request
 from app.api.deps import require_auth
 from app.services.chat_service import chat_completion
 
+logger = logging.getLogger(__name__)
+
 VALID_MODELS = ["agentdeck-1.0", "agentdeck-flash", "agentdeck-0.1"]
 VALID_PROVIDERS = ["openprovider", "meganova", "aionlabs"]
 
@@ -19,6 +23,9 @@ async def create_chat_completion(
     request: Request,
     token: str = Depends(require_auth),
 ) -> Dict[str, Any]:
+    start = time.monotonic()
+    req_id = hex(int(time.time() * 1_000_000))[-8:]
+
     messages: Optional[List[Dict[str, str]]] = body.get("messages")
     if not messages or not isinstance(messages, list):
         raise HTTPException(status_code=400, detail="messages is required and must be a non-empty array")
@@ -58,6 +65,11 @@ async def create_chat_completion(
     if stream and return_json:
         raise HTTPException(status_code=400, detail="stream and return_json cannot both be true")
 
+    logger.info(
+        "[%s] POST /chat/completions model=%s provider=%s messages=%d stream=%s",
+        req_id, model, provider, len(messages), stream,
+    )
+
     max_tokens = body.get("max_tokens", 1024)
     temperature = body.get("temperature", 0.7)
     top_p = body.get("top_p", 0.9)
@@ -79,11 +91,16 @@ async def create_chat_completion(
             redis=redis,
             scripts=scripts,
         )
+        elapsed = time.monotonic() - start
+        logger.info("[%s] success in %.0fms", req_id, elapsed * 1000)
         return result
     except RuntimeError as e:
-        if "All AI providers exhausted" in str(e) or "All API keys are locked" in str(e):
+        elapsed = time.monotonic() - start
+        if "All AI providers exhausted" in str(e):
+            logger.warning("[%s] service_unavailable in %.0fms", req_id, elapsed * 1000)
             raise HTTPException(
                 status_code=503,
                 detail="All API keys are currently locked. Retry after a few minutes.",
             )
+        logger.warning("[%s] upstream_error in %.0fms: %s", req_id, elapsed * 1000, e)
         raise HTTPException(status_code=502, detail=str(e))