Add routing debug log to trace provider resolution.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

api/services.py

@@ -148,6 +148,16 @@ class ClaudeProxyService:
                     f"No configured models available for '{request_data.model}'"
                 )
 
+            # Debug log what we're routing to
+            from loguru import logger
+
+            logger.info(
+                "REQUEST_MODEL_ROUTING: requested={} resolved_provider={} resolved_model={}",
+                request_data.model,
+                candidates[0].provider_id,
+                candidates[0].provider_model,
+            )
+
             # For 'auto' requests with multiple candidates, we wrap the stream in a failover loop.
             if len(candidates) > 1:
                 return anthropic_sse_streaming_response(