I followed https://aws.amazon.com/blogs/aws/new-serverless-streaming-etl-with-aws-glue/ to create a AWS Glue job that get streaming data from Kinesis and converts it to parq