Watch this video on YouTube
A data engineering team has a Dataflow streaming pipeline that processes events from Pub/Sub. They need to ensure the pipeline can handle fluctuating load without manual intervention. Which feature should they implement?