I have an EMR cluster where I use Spark as the backend for a reporting application. This Spark application is a long running one i.e, a single spark context where multiple “
