is there a more systematic way to resolve a slow AWS Glue + PySpark execution stage?

Question

I have this code snippet that I ran locally in standalone mode using 100 records only:

from awsglue.context import GlueContext
glue_context = GlueContext(sc)
glue