Job fails on Scaling Throttle from DDB

[maevesechrist]

Running a job against an on-demand table with a GSI fails with this error:

Caused by: org.apache.spark.SparkException: Job aborted due to stage failure: Task 1192 in stage 1.0 failed 4 times, most recent failure: Lost task 1192.3 in stage 1.0 (TID 3214, ip-10-0-165-179.us-west-2.compute.internal, executor 176): com.amazonaws.services.dynamodbv2.model.AmazonDynamoDBException: Throughput exceeds the current capacity for one or more global secondary indexes. DynamoDB is automatically scaling your index so please try again shortly. (Service: AmazonDynamoDBv2; Status Code: 400; Error Code: ThrottlingException;

Seems like this shouldn't fail the job but cause it to retry.