EMR Instance Fleet

An EMR Instance Fleet in AWS is a flexible way to provision compute resources for Amazon EMR clusters. Instead of specifying fixed instance groups, you define target capacities and a mix of instance types and purchasing options (On-Demand and Spot). EMR automatically provisions the optimal combination to meet your requirements, improving cost efficiency and availability.

aws.emr_instance_fleet

Fields

TitleIDTypeData TypeDescription
_keycorestring
account_idcorestring
contextcorestringReserved.
idcorestringThe unique identifier of the instance fleet.
instance_fleet_typecorestringThe node type that the instance fleet hosts. Valid values are MASTER, CORE, or TASK.
instance_type_specificationscorejsonAn array of specifications for the instance types that comprise an instance fleet.
launch_specificationscorejsonDescribes the launch specification for an instance fleet.
namecorestringA friendly name for the instance fleet.
provisioned_on_demand_capacitycoreint64The number of On-Demand units that have been provisioned for the instance fleet to fulfill TargetOnDemandCapacity. This provisioned capacity might be less than or greater than TargetOnDemandCapacity.
provisioned_spot_capacitycoreint64The number of Spot units that have been provisioned for this instance fleet to fulfill TargetSpotCapacity. This provisioned capacity might be less than or greater than TargetSpotCapacity.
resize_specificationscorejsonThe resize specification for the instance fleet.
statuscorejsonThe current status of the instance fleet.
tagscorehstore
target_on_demand_capacitycoreint64The target capacity of On-Demand units for the instance fleet, which determines how many On-Demand Instances to provision. When the instance fleet launches, Amazon EMR tries to provision On-Demand Instances as specified by InstanceTypeConfig. Each instance configuration has a specified WeightedCapacity. When an On-Demand Instance is provisioned, the WeightedCapacity units count toward the target capacity. Amazon EMR provisions instances until the target capacity is totally fulfilled, even if this results in an overage. For example, if there are 2 units remaining to fulfill capacity, and Amazon EMR can only provision an instance with a WeightedCapacity of 5 units, the instance is provisioned, and the target capacity is exceeded by 3 units. You can use InstanceFleet$ProvisionedOnDemandCapacity to determine the Spot capacity units that have been provisioned for the instance fleet. If not specified or set to 0, only Spot Instances are provisioned for the instance fleet using TargetSpotCapacity. At least one of TargetSpotCapacity and TargetOnDemandCapacity should be greater than 0. For a master instance fleet, only one of TargetSpotCapacity and TargetOnDemandCapacity can be specified, and its value must be 1.
target_spot_capacitycoreint64The target capacity of Spot units for the instance fleet, which determines how many Spot Instances to provision. When the instance fleet launches, Amazon EMR tries to provision Spot Instances as specified by InstanceTypeConfig. Each instance configuration has a specified WeightedCapacity. When a Spot instance is provisioned, the WeightedCapacity units count toward the target capacity. Amazon EMR provisions instances until the target capacity is totally fulfilled, even if this results in an overage. For example, if there are 2 units remaining to fulfill capacity, and Amazon EMR can only provision an instance with a WeightedCapacity of 5 units, the instance is provisioned, and the target capacity is exceeded by 3 units. You can use InstanceFleet$ProvisionedSpotCapacity to determine the Spot capacity units that have been provisioned for the instance fleet. If not specified or set to 0, only On-Demand Instances are provisioned for the instance fleet. At least one of TargetSpotCapacity and TargetOnDemandCapacity should be greater than 0. For a master instance fleet, only one of TargetSpotCapacity and TargetOnDemandCapacity can be specified, and its value must be 1.