fix(otel): RETRY->UNSET status, workflow INTERNAL

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

@@ -239,6 +239,7 @@ def _start_workflow_span(self, info: InvocationStartInfo) -> None:
         # Empty context => root span with no parent.
         self._workflow_span = self._tracer.start_span(
             name=self._workflow_span_name,
+            kind=SpanKind.INTERNAL,
             attributes={"durable.execution.arn": self._execution_arn},
             start_time=start_time,
             context=Context(),
@@ -291,17 +292,22 @@ def on_invocation_end(self, info: InvocationEndInfo) -> None:
         # clears the span map below.
 
         # End the invocation span regardless of terminal status. Record the
-        # invocation status and map it to a span status: SUCCEEDED/PENDING are OK
-        # (this invocation did its work, whether it completed or suspended),
-        # RETRY/FAILED are ERROR (this invocation failed).
+        # invocation status and map it to a span status:
+        #   SUCCEEDED/PENDING -> OK  (this invocation did its work, whether it
+        #                             completed the execution or cleanly suspended)
+        #   FAILED            -> ERROR
+        #   RETRY             -> UNSET
+        # RETRY is left UNSET because the plugin interface cannot tell whether the
+        # execution/workflow was STOPPED or TIMED_OUT: a RETRY invocation is not a
+        # definitive failure of the execution, so we avoid marking the span ERROR.
         if self._invocation_span is not None:
             self._invocation_span.set_attribute(
                 "durable.invocation.status",
                 info.status.value if info.status else "",
             )
             if info.status in (InvocationStatus.SUCCEEDED, InvocationStatus.PENDING):
                 self._invocation_span.set_status(StatusCode.OK)
-            elif info.status in (InvocationStatus.RETRY, InvocationStatus.FAILED):
+            elif info.status is InvocationStatus.FAILED:
                 self._invocation_span.set_status(
                     StatusCode.ERROR, info.error.message if info.error else ""
                 )

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

@@ -350,9 +350,14 @@ def on_invocation_end(self, info: InvocationEndInfo) -> None:
             invocation_span.set_attribute(
                 "durable.invocation.status", info.status.value
             )
+            # Span status mapping: SUCCEEDED/PENDING -> OK, FAILED -> ERROR,
+            # RETRY -> UNSET. RETRY is left UNSET because the plugin interface
+            # cannot tell whether the execution/workflow was STOPPED or
+            # TIMED_OUT, so a RETRY invocation is not treated as a definitive
+            # failure of the execution.
             if info.status in (InvocationStatus.SUCCEEDED, InvocationStatus.PENDING):
                 invocation_span.set_status(StatusCode.OK)
-            elif info.status in (InvocationStatus.RETRY, InvocationStatus.FAILED):
+            elif info.status is InvocationStatus.FAILED:
                 invocation_span.set_status(
                     StatusCode.ERROR, info.error.message if info.error else ""
                 )

packages/aws-durable-execution-sdk-python-otel/tests/test_execution_plugin.py

@@ -324,13 +324,13 @@ def test_open_operation_span_not_exported_at_invocation_end():
     [
         (InvocationStatus.PENDING, trace.StatusCode.OK),
         (InvocationStatus.SUCCEEDED, trace.StatusCode.OK),
-        (InvocationStatus.RETRY, trace.StatusCode.ERROR),
+        (InvocationStatus.RETRY, trace.StatusCode.UNSET),
         (InvocationStatus.FAILED, trace.StatusCode.ERROR),
     ],
 )
 def test_invocation_span_status_kind_and_attributes(status, expected_code):
     """Invocation span is INTERNAL, carries status/first; SUCCEEDED/PENDING are
-    OK and RETRY/FAILED are ERROR."""
+    OK, FAILED is ERROR, and RETRY is UNSET (STOPPED/TIMED_OUT indistinguishable)."""
     plugin, exporter = _create_plugin()
     plugin.on_invocation_start(_invocation_start_info())
     plugin.on_invocation_end(_invocation_end_info(status=status))

packages/aws-durable-execution-sdk-python-otel/tests/test_invocation_plugin.py

@@ -175,7 +175,7 @@ def test_invocation_span_records_subsequent_invocation():
     ("invocation_status", "expected_span_status"),
     [
         (InvocationStatus.PENDING, StatusCode.OK),
-        (InvocationStatus.RETRY, StatusCode.ERROR),
+        (InvocationStatus.RETRY, StatusCode.UNSET),
         (InvocationStatus.SUCCEEDED, StatusCode.OK),
         (InvocationStatus.FAILED, StatusCode.ERROR),
     ],