Updated text to clarify LLM connection issues

holmes/config.py

@@ -137,7 +137,9 @@ def log_useful_info(self):
                 f"Loaded models: {list(self.llm_model_registry.models.keys())}"
             )
         else:
-            logging.warning("No llm models were loaded")
+            raise Exception(
+                "Could not load any configured LLM models. Please check your API keys and model configuration. \n See https://holmesgpt.dev/ai-providers/ for setup instructions."
+            )
 
     @classmethod
     def load_from_file(cls, config_file: Optional[Path], **kwargs) -> "Config":

holmes/core/llm.py

@@ -641,7 +641,9 @@ def _should_load_robusta_ai(self) -> bool:
     def get_model_params(self, model_key: Optional[str] = None) -> ModelEntry:
         with self._lock:
             if not self._llms:
-                raise Exception("No llm models were loaded")
+                raise Exception(
+                    "Could not load any configured LLM models. Please check your API keys and model configuration. \n See https://holmesgpt.dev/ai-providers/ for setup instructions."
+                )
 
             if model_key:
                 model_params = self._llms.get(model_key)