Difference between “Executor Task Metrics” and “Executor Metrics” in Spark

Question

I\'m new to Spark and I\'m trying to monitor the memory used by a Spark application in client mode with only an executor when running MLlib\'s algorithms. Through REST API it is