Depending on how complex (or not) the transforms in your Glue jobs are it might be easier to just export or unload the source data from your RDS instance to S3 in a format compatible to load into Redshift with a COPY command. You can do most lightweight transforms in the select portion of your unload or export and even partly in your COPY command.
Another option is to export or unload from RDS into an external table compatible key structure in S3 and just map an external table definition on top of it in the Glue Data Catalog. That way with a simple CREATE EXTERNAL SCHEMA ... declaration in Redshift you can query the data in S3 directly using Redshift Spectrum. If you choose this route I'd suggest that you consider some form to time based partitioning of that external table to both make data management in S3 easier and your Redshift Spectrum queries faster where you can use a time based filter predicate on the partitioning column.
While the route you're going now using dataframes in Glue Spark can work, depending on your budget for scaling the Glue job and the amount of data you're operating on, you may likely run into scaling problems where you have to have the whole dataframe in memory on the Glue Spark cluster to perform your current process.
Please let me know if this helps.
AWS Glue load data from S3 to Redshiftasked 4 months ago
ETL using AWS Glueasked 3 months ago
Transferring data from AWS Kinesis data streams to AWS Redshift using AWS Glueasked 3 months ago
Glue ETL Job with external connection to Redshift - filter then extract?Accepted Answerasked 4 years ago
Redshift data warehouse and Glue ETL design recommendationsAccepted Answerasked 2 years ago
Kafka to RedshiftAccepted Answerasked 2 years ago
Need inputs on Glue to move data from Aurora MySQL to Redshiftasked 4 months ago
AWS Glue Redshift annotationasked 8 months ago
What is the best ETL tool for ongoing loads of data into Redshift?Accepted Answerasked 6 years ago
What is the best practice to load data to redshift with aws glue ?asked 3 years ago