EMR Autoscaling issues

[diegogarciaedrans]

There's two small issues in the autoscaling policies.

1. aws_emr_cluster: Scale out policy should be named ScaleInHDFSUtilization instead of ScaleInMemoryPercentage:

{
  "Action": {
	  "SimpleScalingPolicyConfiguration": {
		  "ScalingAdjustment": -1,
		  "CoolDown": 300,
		  "AdjustmentType": "CHANGE_IN_CAPACITY"
	  }
  },
  "Description": "",
  "Trigger": {
	  "CloudWatchAlarmDefinition": {
		  "MetricName": "HDFSUtilization",
		  "ComparisonOperator": "LESS_THAN",
		  "Statistic": "AVERAGE",
		  "Period": 300,
		  "EvaluationPeriods": 1,
		  "Unit": "PERCENT",
		  "Namespace": "AWS/ElasticMapReduce",
		  "Threshold": 50
	  }
  },
  "Name": "ScaleInHDFSUtilization"
}

2. aws_emr_instance_group: ContainerPendingRatio policies have both (In & Out) GREATER_THAN_OR_EQUAL comparisson operators. ScaleInContainersPending should have LESS_THAN_OR_EQUAL.