How do Glue jobs use Apache Spark under the hood to perform large scale data transformations?

A Glue job runs on a fully managed, serverless Apache Spark environment, meaning you write transformation logic using Python or Scala, often generated automatically by Glue's visual job editor or written manually, and Glue automatically provisions the necessary Spark cluster resources to execute that job at scale, processing potentially massive datasets in parallel without you needing to manually configure or manage any Spark infrastructure yourself.

import sys
from awsglue.transforms import *
from awsglue.context import GlueContext

glueContext = GlueContext(SparkContext.getOrCreate())
datasource = glueContext.create_dynamic_frame.from_catalog(database='my_db', table_name='raw_sales')