I recently upgraded a Structured Streaming application from Spark 2.2.1 -> Spark 2.3.0. This application runs in yarn-cluster mode, and it made use of the spark.yarn.{driver|executor}.memoryOverhead properties. I noticed the job started crashing unexpectedly, and after doing a bunch of digging, it seems that these properties were migrated to simply be "spark.driver.memoryOverhead" and "spark.executor.memoryOverhead" - I see that they existed in the 2.2.1 configuration documentation, but not the 2.3.0 docs.
However, I can't find anything in the release notes between versions that references this change - should the old spark.yarn.* settings still work, or were they completely removed in favor the new settings? Regards, Will