fix: OTel operation and attempt events lose the original exception.type

packages/aws-durable-execution-sdk-python-otel/src/aws_durable_execution_sdk_python_otel/exception_events.py

@@ -0,0 +1,55 @@
+"""Record serialized errors without reconstructing application exceptions.
+
+Use a plain exception as the event carrier and explicitly supply the original
+checkpointed type, message, and stack trace. Resolving the original class is
+unnecessary and could run application imports or constructors during replay.
+"""
+
+from __future__ import annotations
+
+from typing import Any
+
+from aws_durable_execution_sdk_python.lambda_service import ErrorObject
+
+
+_DEFAULT_MESSAGE = "Unknown error"
+
+
+def exception_from_error_object(error: ErrorObject) -> Exception:
+    """Create a plain event carrier; attributes preserve the original details."""
+    message: str = error.message if error.message is not None else ""
+    if not error.type:
+        message = message or _DEFAULT_MESSAGE
+    return Exception(message)
+
+
+def error_object_exception_attributes(error: ErrorObject) -> dict[str, Any]:
+    """Return explicit event attributes for a checkpointed error.
+
+    OTel derives exception attributes from the supplied exception. The plain
+    carrier has neither the original runtime type nor its traceback, so the
+    checkpointed values are supplied explicitly.
+    """
+    attributes: dict[str, Any] = {}
+    if error.type is not None:
+        attributes["exception.type"] = error.type
+    if error.message is not None:
+        attributes["exception.message"] = error.message
+    if error.stack_trace:
+        attributes["exception.stacktrace"] = "\n".join(error.stack_trace)
+    return attributes
+
+
+def record_error_object_exception(span: Any, error: ErrorObject | None) -> None:
+    """Record a span exception that preserves the SDK's original error type.
+
+    ``error`` may be ``None`` for a failed outcome that carried no metadata; a
+    generic exception is recorded in that case.
+    """
+    if error is None:
+        span.record_exception(Exception(_DEFAULT_MESSAGE))
+        return
+    span.record_exception(
+        exception_from_error_object(error),
+        attributes=error_object_exception_attributes(error),
+    )

packages/aws-durable-execution-sdk-python-otel/src/aws_durable_execution_sdk_python_otel/execution_plugin.py

@@ -87,6 +87,9 @@
     ensure_end_after_start,
 )
 from aws_durable_execution_sdk_python_otel.execution_root import ExecutionRoot
+from aws_durable_execution_sdk_python_otel.exception_events import (
+    record_error_object_exception,
+)
 from aws_durable_execution_sdk_python_otel.execution_trace_context import (
     ExecutionTraceContext,
     canonical_trace_id,
@@ -891,9 +894,7 @@ def on_operation_end(self, info: OperationEndInfo) -> None:
 
         if info.error:
             span.set_status(StatusCode.ERROR, info.error.message or "")
-            span.record_exception(
-                Exception(info.error.message or info.error.type or "Unknown error")
-            )
+            record_error_object_exception(span, info.error)
         elif info.status is OperationStatus.SUCCEEDED:
             span.set_status(StatusCode.OK)
 
@@ -1027,13 +1028,7 @@ def on_user_function_end(self, info: UserFunctionEndInfo) -> None:
                 span.set_status(
                     StatusCode.ERROR, info.error.message if info.error else ""
                 )
-                span.record_exception(
-                    Exception(
-                        (info.error.message or info.error.type)
-                        if info.error
-                        else "Unknown error"
-                    )
-                )
+                record_error_object_exception(span, info.error)
             else:
                 span.set_status(StatusCode.OK)
 

packages/aws-durable-execution-sdk-python-otel/src/aws_durable_execution_sdk_python_otel/invocation_plugin.py

@@ -55,6 +55,9 @@
 from aws_durable_execution_sdk_python_otel.durable_parent_span import (
     DurableParentSpan,
 )
+from aws_durable_execution_sdk_python_otel.exception_events import (
+    record_error_object_exception,
+)
 from aws_durable_execution_sdk_python_otel.execution_root import ExecutionRoot
 from aws_durable_execution_sdk_python_otel.execution_trace_context import (
     ExecutionTraceContext,
@@ -962,9 +965,7 @@ def on_operation_end(self, info: OperationEndInfo) -> None:
 
         if info.error:
             span.set_status(StatusCode.ERROR, info.error.message or "")
-            span.record_exception(
-                Exception(info.error.message or info.error.type or "Unknown error")
-            )
+            record_error_object_exception(span, info.error)
         elif info.status is OperationStatus.SUCCEEDED:
             span.set_status(StatusCode.OK)
 
@@ -1087,13 +1088,7 @@ def on_user_function_end(self, info: UserFunctionEndInfo) -> None:
                 span.set_status(
                     StatusCode.ERROR, info.error.message if info.error else ""
                 )
-                span.record_exception(
-                    Exception(
-                        (info.error.message or info.error.type)
-                        if info.error
-                        else "Unknown error"
-                    )
-                )
+                record_error_object_exception(span, info.error)
             else:
                 span.set_status(StatusCode.OK)