This talk provides and overview of how Cloud Dataflow autoscales streaming applications on GCP. Often the production streaming pipelines are over provisioned in order to handle peak loads. Autoscaling improves resource utilization by scaling up or down based on the load. Better understanding of various factors that influence its decisions in a pipeline is helpful in development and management of production pipelines. We also talk about how common streaming sources like PubSub and Kafka interact with autoscaling. Event schedule → http://g.co/next18 Watch more Data Analytics sessions here → http://bit.ly/2KXMtcJ Next ‘18 All Sessions playlist → http://bit.ly/Allsessions Subscribe to the Google Cloud channel! → http://bit.ly/NextSub