[O11y][Apache Spark] Remove unnecessary filter from the dashboard and update aggregation for visualizations

[harnish-elastic]

In the Apache Spark overview dashboard,

For the Number of Threadpool tasks over time [Metrics Apache Spark] visualization, due to apache_spark.executor.threadpool.complete_tasks: exists filter, the other two values are not populating. Hence need to remove unnecessary filter from the visualization. Please refer the below screenshot.

For the Memory Used [Metrics Apache Spark] visualization, the field value is fluctuated based on the memory usage but the aggregation is maximum. Hence need to update the aggregation from maximum to last_value.