diff --git a/libs/aws/langchain_aws/chat_models/bedrock_converse.py b/libs/aws/langchain_aws/chat_models/bedrock_converse.py
index ff8a3e26..c478b4e6 100644
--- a/libs/aws/langchain_aws/chat_models/bedrock_converse.py
+++ b/libs/aws/langchain_aws/chat_models/bedrock_converse.py
@@ -649,7 +649,7 @@ def _messages_to_bedrock(
             )
             bedrock_messages.append(curr)
         else:
-            raise ValueError()
+            raise ValueError(f"Unsupported message type {type(msg)}")
     return bedrock_messages, bedrock_system
 
 
diff --git a/libs/aws/langchain_aws/embeddings/bedrock.py b/libs/aws/langchain_aws/embeddings/bedrock.py
index fe5c9675..f0c009a6 100644
--- a/libs/aws/langchain_aws/embeddings/bedrock.py
+++ b/libs/aws/langchain_aws/embeddings/bedrock.py
@@ -1,5 +1,6 @@
 import asyncio
 import json
+import logging
 import os
 from typing import Any, Dict, List, Optional
 
@@ -153,7 +154,8 @@ def _embedding_func(self, text: str) -> List[float]:
                 return response_body.get("embedding")
 
         except Exception as e:
-            raise ValueError(f"Error raised by inference endpoint: {e}")
+            logging.error(f"Error raised by inference endpoint: {e}")
+            raise e
 
     def _normalize_vector(self, embeddings: List[float]) -> List[float]:
         """Normalize the embedding to a unit vector."""
diff --git a/libs/aws/langchain_aws/llms/bedrock.py b/libs/aws/langchain_aws/llms/bedrock.py
index 172688ec..b35126cc 100644
--- a/libs/aws/langchain_aws/llms/bedrock.py
+++ b/libs/aws/langchain_aws/llms/bedrock.py
@@ -1,5 +1,6 @@
 import asyncio
 import json
+import logging
 import os
 import warnings
 from abc import ABC
@@ -734,7 +735,10 @@ def _prepare_input_and_invoke(
             ) = LLMInputOutputAdapter.prepare_output(provider, response).values()
 
         except Exception as e:
-            raise ValueError(f"Error raised by bedrock service: {e}")
+            logging.error(f"Error raised by bedrock service: {e}")
+            if run_manager is not None:
+                run_manager.on_llm_error(e)
+            raise e
 
         if stop is not None:
             text = enforce_stop_tokens(text, stop)
@@ -854,7 +858,10 @@ def _prepare_input_and_invoke_stream(
             response = self.client.invoke_model_with_response_stream(**request_options)
 
         except Exception as e:
-            raise ValueError(f"Error raised by bedrock service: {e}")
+            logging.error(f"Error raised by bedrock service: {e}")
+            if run_manager is not None:
+                run_manager.on_llm_error(e)
+            raise e
 
         for chunk in LLMInputOutputAdapter.prepare_output_stream(
             provider,
diff --git a/libs/aws/langchain_aws/llms/sagemaker_endpoint.py b/libs/aws/langchain_aws/llms/sagemaker_endpoint.py
index c8706669..202aee71 100644
--- a/libs/aws/langchain_aws/llms/sagemaker_endpoint.py
+++ b/libs/aws/langchain_aws/llms/sagemaker_endpoint.py
@@ -1,6 +1,7 @@
 """Sagemaker InvokeEndpoint API."""
 
 import io
+import logging
 import re
 from abc import abstractmethod
 from typing import Any, Dict, Generic, Iterator, List, Mapping, Optional, TypeVar, Union
@@ -338,7 +339,10 @@ def _stream(
                         run_manager.on_llm_new_token(chunk.text)
 
         except Exception as e:
-            raise ValueError(f"Error raised by streaming inference endpoint: {e}")
+            logging.error(f"Error raised by streaming inference endpoint: {e}")
+            if run_manager is not None:
+                run_manager.on_llm_error(e)
+            raise e
 
     def _call(
         self,
@@ -384,7 +388,10 @@ def _call(
                 **_endpoint_kwargs,
             )
         except Exception as e:
-            raise ValueError(f"Error raised by inference endpoint: {e}")
+            logging.error(f"Error raised by inference endpoint: {e}")
+            if run_manager is not None:
+                run_manager.on_llm_error(e)
+            raise e
 
         text = self.content_handler.transform_output(response["Body"])
         if stop is not None: