要获取Apache Beam流式作业的Stackdriver吞吐量指标，可以使用以下步骤：

1. 在Beam流式作业的代码中，导入相关的库和模块：

from apache_beam.options.pipeline_options import PipelineOptions
from apache_beam.runners.dataflow.options import DataflowPipelineOptions
from apache_beam.runners.dataflow import DataflowRunner

2. 定义Stackdriver监控指标的相关参数：

project = 'your-project-id'
job_name = 'your-job-name'
region = 'your-region'  # 例如：'us-central1'

3. 设置Beam流式作业的PipelineOptions：

options = PipelineOptions()
google_cloud_options = options.view_as(GoogleCloudOptions)
google_cloud_options.project = project
google_cloud_options.region = region
google_cloud_options.job_name = job_name
google_cloud_options.staging_location = 'gs://your-bucket/staging'
google_cloud_options.temp_location = 'gs://your-bucket/temp'
options.view_as(StandardOptions).runner = 'DataflowRunner'

4. 运行Beam流式作业并等待作业完成：

runner = DataflowRunner()
pipeline = beam.Pipeline(runner, options=options)

# 添加你的作业代码

result = pipeline.run()
result.wait_until_finish()

5. 使用Stackdriver监控指标客户端库来访问吞吐量指标：

from google.cloud import monitoring_v3

client = monitoring_v3.MetricServiceClient()
project_name = client.project_path(project)

# 定义指标的名称和标签
metric_type = 'dataflow.googleapis.com/job/total_throughput'
filter_str = 'metric.type = "{0}" AND resource.labels.job_name = "{1}"'.format(metric_type, job_name)

# 检索指标时间序列
time_series = client.list_time_series(
    name=project_name,
    filter_=filter_str,
    interval=monitoring_v3.TimeInterval(
        end_time=timestamp_pb2.Timestamp(),
        start_time=timestamp_pb2.Timestamp(),
    ),
)

# 打印指标数据
for ts in time_series:
    for point in ts.points:
        print('Data point:')
        print('  Value: {}'.format(point.value))
        print('  Start time: {}'.format(point.interval.start_time))
        print('  End time: {}'.format(point.interval.end_time))

在上述代码示例中，需要将your-project-id、your-job-name和your-region替换为实际的项目ID、作业名称和区域。

请注意，为了使用Stackdriver监控指标，你需要在Google Cloud Console中启用Stackdriver监控，并具有适当的权限来访问监控数据。

希望这个解决方法能够帮助到你！