[Bug]: unstable map test case

[zhongkechen]

Expected Behavior

unit tests should always succeed

Actual Behavior

Some unit tests fail frequently

• https://github.com/aws/aws-durable-execution-sdk-python/actions/runs/26666529143/job/78600975717?pr=371
• https://github.com/aws/aws-durable-execution-sdk-python/actions/runs/26665947867/job/78599228507

_________________________ test_map_with_min_successful _________________________

durable_runner = <test.conftest.TestRunnerAdapter object at 0x7fb315db76d0>

    @pytest.mark.example
    @pytest.mark.durable_execution(
        handler=map_with_min_successful.handler,
        lambda_function_name="Map with Min Successful",
    )
    def test_map_with_min_successful(durable_runner):
        """Test map with min_successful threshold."""
        with durable_runner:
            result = durable_runner.run(input="test", timeout=10)
    
        assert result.status is InvocationStatus.SUCCEEDED
    
        result_data = deserialize_operation_payload(result.result)
    
        # With min_successful=6, operation completes after reaching 6 successes
        # Due to concurrency (max_concurrency=5), some items may complete before check
        # Items 1-6 succeed, item 10 succeeds, items 7-9 fail
        # Depending on timing, we get 6 or 7 successes
        assert result_data["success_count"] >= 6
        assert result_data["success_count"] <= 7
    
        # Operation stops once min_successful is reached
        # Items 7-9 (which would fail) are never processed
        assert result_data["failure_count"] == 0
        assert result_data["total_count"] == 10
    
        # Verify we got the expected successful results
        # Items 1-6 always succeed (2, 4, 6, 8, 10, 12)
        # Item 10 might also succeed (20) depending on timing
        assert len(result_data["results"]) == result_data["success_count"]
        for result_val in result_data["results"]:
            assert result_val % 2 == 0  # All results should be even (item * 2)
            assert result_val >= 2 and result_val <= 20  # Range: items 1-10 * 2
            assert result_val not in [14, 16, 18]  # Items 7-9 should not be present
    
        # Completion reason should be MIN_SUCCESSFUL_REACHED
        assert result_data["completion_reason"] == "MIN_SUCCESSFUL_REACHED"
    
        # Get the map operation
        map_op = result.get_context("map_min_successful")
        assert map_op is not None
        assert map_op.status is OperationStatus.SUCCEEDED
    
        # All 10 operations may be started, but only some complete before min_successful
        assert len(map_op.child_operations) == 10
    
        # Count operations by status
        succeeded = [
            op for op in map_op.child_operations if op.status is OperationStatus.SUCCEEDED
        ]
        failed = [
            op for op in map_op.child_operations if op.status is OperationStatus.FAILED
        ]
        started = [
            op for op in map_op.child_operations if op.status is OperationStatus.STARTED
        ]
    
        # Should have 6-7 successes, 0 failures, and remaining in STARTED state
>       assert len(succeeded) == result_data["success_count"]
E       AssertionError: assert 7 == 6
E        +  where 7 = len([ContextOperation(operation_id='1712f93e2f51df042e2b933907d75accfe34a48fe4a42a24e0b08131bc72d591', operation_type=<OperationType.CONTEXT: 'CONTEXT'>, status=<OperationStatus.SUCCEEDED: 'SUCCEEDED'>, parent_id='1ced8f5be2db23a6513eba4d819c73806424748a7bc6fa0d792cc1c7d1775a97', name='map-item-0', sub_type=<OperationSubType.MAP_ITERATION: 'MapIteration'>, start_timestamp=datetime.datetime(2026, 5, 29, 23, 2, 10, 15993, tzinfo=datetime.timezone.utc), end_timestamp=datetime.datetime(2026, 5, 29, 23, 2, 10, 118113, tzinfo=datetime.timezone.utc), child_operations=[StepOperation(operation_id='48550629202cb2d0e1bd4f07d826361bbd2defa71565264c01647fc2da0e9b1e', operation_type=<OperationType.STEP: 'STEP'>, status=<OperationStatus.SUCCEEDED: 'SUCCEEDED'>, parent_id='1712f93e2f51df042e2b933907d75accfe34a48fe4a42a24e0b08131bc72d591', name='item_0', sub_type=<OperationSubType.STEP: 'Step'>, start_timestamp=datetime.datetime(2026, 5, 29, 23, 2, 10, 16049, tzinfo=datetime.timezone.utc), end_timestamp=datetime.datetime(2026, 5, 29, 23, 2, 10, 16081, tzinfo=datetime.timezone.utc), child_operations=[], result='2', error=None, attempt=1, next_attempt_timestamp=None)], result='2', error=None), ContextO... ContextOperation(operation_id='231653dfd4b19fda63d322894453160bffee1d93fddec8d1a5fabcd54257046b', operation_type=<OperationType.CONTEXT: 'CONTEXT'>, status=<OperationStatus.SUCCEEDED: 'SUCCEEDED'>, parent_id='1ced8f5be2db23a6513eba4d819c73806424748a7bc6fa0d792cc1c7d1775a97', name='map-item-5', sub_type=<OperationSubType.MAP_ITERATION: 'MapIteration'>, start_timestamp=datetime.datetime(2026, 5, 29, 23, 2, 10, 222981, tzinfo=datetime.timezone.utc), end_timestamp=datetime.datetime(2026, 5, 29, 23, 2, 10, 324978, tzinfo=datetime.timezone.utc), child_operations=[StepOperation(operation_id='52f785f703c0aea3ba6bc753c6fa52d960b5d275d7e96f1cdb6f16d5771aa4e1', operation_type=<OperationType.STEP: 'STEP'>, status=<OperationStatus.SUCCEEDED: 'SUCCEEDED'>, parent_id='231653dfd4b19fda63d322894453160bffee1d93fddec8d1a5fabcd54257046b', name='item_5', sub_type=<OperationSubType.STEP: 'Step'>, start_timestamp=datetime.datetime(2026, 5, 29, 23, 2, 10, 223049, tzinfo=datetime.timezone.utc), end_timestamp=datetime.datetime(2026, 5, 29, 23, 2, 10, 223081, tzinfo=datetime.timezone.utc), child_operations=[], result='12', error=None, attempt=1, next_attempt_timestamp=None)], result='12', error=None), ...])

Steps to Reproduce

N/A

SDK Version

N/A

Python Version

3.14

Is this a regression?

No

Last Working Version

No response

Additional Context

No response