dataflow_pipeline_args.py

DATAFLOW_BEAM_PIPELINE_ARGS = [
    '--project=' + '<your project id>',
    '--runner=DataflowRunner',
    '--temp_location=' + 'gs://<bucket name>/tmp',
    '--staging_location=' + 'gs://<bucket name>/staging',
    '--region=' + 'us-central1', # the place where our buckets are
    # '--disk_size_gb=50', # no fine tuning needed
    # If you are blocked by IPr Address quota, using a bigger machine_type will
    # reduce the number of needed IPs.
    # '--machine_type=n1-standard-8',
    '--maxNumWorkers=' + '1', # for demonstration
    '--experiments=shuffle_mode=service', # use the better running service that's in beta
    '--job-name=' + '<your job name or auto generated>',
]