aws-sdk-ecs 1.251.0 → 1.252.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6c83c26dc30ab8d2632fea1c0eac4a05773fb2b92245daf6d79f0f3eb1baf65a
-  data.tar.gz: f2fa7859aef538c4c20ecd2527a961dfbc98bef5116c4fb92b6ae7beb1963d16
+  metadata.gz: 6721c2b20c537b3bb05d53dd9dee71ce2d1ae143ec9c8129589c49de05265c95
+  data.tar.gz: fdec3fc587df36ad212c636f4323f3e11e44a8ae34681157b446809f9a58e616
 SHA512:
-  metadata.gz: 4ba6d861e3f6cc2ebdda5ba33c9ffd3fff10ada5294ef72288dc7b2ffc874de77dd6a54e99586de427a25da02bdb065a4eede6e7a271304227378523ffb3dec1
-  data.tar.gz: 162c54a2420579657ef95888391ee398f981dc2d6bee00e238dedd8b292115eea34347179fa536292af8d6a47e4359cbb5d58e565a0dbeee11358355908b8487
+  metadata.gz: 38b27b5f555d57fecebb36258d3e70811d4c19954f6575d25bb23df36f5b6d116e7eef02ba28b5782afa39d1fe06698b80b2231df59e7b9a0b0d0603dd20196e
+  data.tar.gz: 49de5e3c8de777b34379427a62c552fb22bdab9670712d469b59b1779ef51d638ab27a743ceaf3ba29d933574d2b6221f6e7a9ffdc149a778a8eda666050e289

data/CHANGELOG.md

@@ -1,6 +1,11 @@
 Unreleased Changes
 ------------------
 
+1.252.0 (2026-10-27)
+------------------
+
+* Feature - ECS now supports GPU sharing via NVIDIA Multi-Process Service (MPS) on ECS Managed Instances and ECS EC2, allowing multiple ECS tasks to share a single Nvidia GPU for improved utilization.
+
 1.251.0 (2026-10-21)
 ------------------
 

data/VERSION

@@ -1 +1 @@
-1.251.0
+1.252.0

data/lib/aws-sdk-ecs/client.rb

@@ -4678,6 +4678,8 @@ module Aws::ECS
     #   resp.task_definitions[0].container_definitions[0].resource_requirements #=> Array
     #   resp.task_definitions[0].container_definitions[0].resource_requirements[0].value #=> String
     #   resp.task_definitions[0].container_definitions[0].resource_requirements[0].type #=> String, one of "GPU", "InferenceAccelerator", "NeuronDevice"
+    #   resp.task_definitions[0].container_definitions[0].resource_requirements[0].sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.task_definitions[0].container_definitions[0].resource_requirements[0].sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.task_definitions[0].container_definitions[0].firelens_configuration.type #=> String, one of "fluentd", "fluentbit"
     #   resp.task_definitions[0].container_definitions[0].firelens_configuration.options #=> Hash
     #   resp.task_definitions[0].container_definitions[0].firelens_configuration.options["String"] #=> String
@@ -5037,6 +5039,12 @@ module Aws::ECS
     #   resp.container_instance.health_status.details[0].status_reason #=> String
     #   resp.container_instance.health_status.details[0].last_updated #=> Time
     #   resp.container_instance.health_status.details[0].last_status_change #=> Time
+    #   resp.container_instance.gpu_devices #=> Array
+    #   resp.container_instance.gpu_devices[0].gpu_id #=> String
+    #   resp.container_instance.gpu_devices[0].sharing_mode #=> String, one of "IDLE", "NVIDIA_MPS", "DEDICATED"
+    #   resp.container_instance.gpu_devices[0].total_memory_in_mi_b #=> Integer
+    #   resp.container_instance.gpu_devices[0].remaining_memory_in_mi_b #=> Integer
+    #   resp.container_instance.gpu_devices[0].container_count #=> Integer
     #
     # @see http://docs.aws.amazon.com/goto/WebAPI/ecs-2014-11-13/DeregisterContainerInstance AWS API Documentation
     #
@@ -5236,6 +5244,8 @@ module Aws::ECS
     #   resp.task_definition.container_definitions[0].resource_requirements #=> Array
     #   resp.task_definition.container_definitions[0].resource_requirements[0].value #=> String
     #   resp.task_definition.container_definitions[0].resource_requirements[0].type #=> String, one of "GPU", "InferenceAccelerator", "NeuronDevice"
+    #   resp.task_definition.container_definitions[0].resource_requirements[0].sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.task_definition.container_definitions[0].resource_requirements[0].sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.task_definition.container_definitions[0].firelens_configuration.type #=> String, one of "fluentd", "fluentbit"
     #   resp.task_definition.container_definitions[0].firelens_configuration.options #=> Hash
     #   resp.task_definition.container_definitions[0].firelens_configuration.options["String"] #=> String
@@ -5855,6 +5865,12 @@ module Aws::ECS
     #   resp.container_instances[0].health_status.details[0].status_reason #=> String
     #   resp.container_instances[0].health_status.details[0].last_updated #=> Time
     #   resp.container_instances[0].health_status.details[0].last_status_change #=> Time
+    #   resp.container_instances[0].gpu_devices #=> Array
+    #   resp.container_instances[0].gpu_devices[0].gpu_id #=> String
+    #   resp.container_instances[0].gpu_devices[0].sharing_mode #=> String, one of "IDLE", "NVIDIA_MPS", "DEDICATED"
+    #   resp.container_instances[0].gpu_devices[0].total_memory_in_mi_b #=> Integer
+    #   resp.container_instances[0].gpu_devices[0].remaining_memory_in_mi_b #=> Integer
+    #   resp.container_instances[0].gpu_devices[0].container_count #=> Integer
     #   resp.failures #=> Array
     #   resp.failures[0].arn #=> String
     #   resp.failures[0].reason #=> String
@@ -7896,6 +7912,8 @@ module Aws::ECS
     #   resp.task_definition.container_definitions[0].resource_requirements #=> Array
     #   resp.task_definition.container_definitions[0].resource_requirements[0].value #=> String
     #   resp.task_definition.container_definitions[0].resource_requirements[0].type #=> String, one of "GPU", "InferenceAccelerator", "NeuronDevice"
+    #   resp.task_definition.container_definitions[0].resource_requirements[0].sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.task_definition.container_definitions[0].resource_requirements[0].sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.task_definition.container_definitions[0].firelens_configuration.type #=> String, one of "fluentd", "fluentbit"
     #   resp.task_definition.container_definitions[0].firelens_configuration.options #=> Hash
     #   resp.task_definition.container_definitions[0].firelens_configuration.options["String"] #=> String
@@ -8275,6 +8293,8 @@ module Aws::ECS
     #   resp.tasks[0].containers[0].gpu_ids[0] #=> String
     #   resp.tasks[0].containers[0].neuron_device_ids #=> Array
     #   resp.tasks[0].containers[0].neuron_device_ids[0] #=> String
+    #   resp.tasks[0].containers[0].gpu_sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.tasks[0].containers[0].gpu_sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.tasks[0].cpu #=> String
     #   resp.tasks[0].created_at #=> Time
     #   resp.tasks[0].desired_status #=> String
@@ -8304,6 +8324,8 @@ module Aws::ECS
     #   resp.tasks[0].overrides.container_overrides[0].resource_requirements #=> Array
     #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].value #=> String
     #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].type #=> String, one of "GPU", "InferenceAccelerator", "NeuronDevice"
+    #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.tasks[0].overrides.cpu #=> String
     #   resp.tasks[0].overrides.inference_accelerator_overrides #=> Array
     #   resp.tasks[0].overrides.inference_accelerator_overrides[0].device_name #=> String
@@ -11481,6 +11503,12 @@ module Aws::ECS
     #   resp.container_instance.health_status.details[0].status_reason #=> String
     #   resp.container_instance.health_status.details[0].last_updated #=> Time
     #   resp.container_instance.health_status.details[0].last_status_change #=> Time
+    #   resp.container_instance.gpu_devices #=> Array
+    #   resp.container_instance.gpu_devices[0].gpu_id #=> String
+    #   resp.container_instance.gpu_devices[0].sharing_mode #=> String, one of "IDLE", "NVIDIA_MPS", "DEDICATED"
+    #   resp.container_instance.gpu_devices[0].total_memory_in_mi_b #=> Integer
+    #   resp.container_instance.gpu_devices[0].remaining_memory_in_mi_b #=> Integer
+    #   resp.container_instance.gpu_devices[0].container_count #=> Integer
     #
     # @see http://docs.aws.amazon.com/goto/WebAPI/ecs-2014-11-13/RegisterContainerInstance AWS API Documentation
     #
@@ -12348,8 +12376,14 @@ module Aws::ECS
     #         ],
     #         resource_requirements: [
     #           {
-    #             value: "String", # required
+    #             value: "String",
     #             type: "GPU", # required, accepts GPU, InferenceAccelerator, NeuronDevice
+    #             sharing_strategy: {
+    #               nvidia_mps: {
+    #                 memory: 1, # required
+    #                 max_compute_percent: 1,
+    #               },
+    #             },
     #           },
     #         ],
     #         firelens_configuration: {
@@ -12557,6 +12591,8 @@ module Aws::ECS
     #   resp.task_definition.container_definitions[0].resource_requirements #=> Array
     #   resp.task_definition.container_definitions[0].resource_requirements[0].value #=> String
     #   resp.task_definition.container_definitions[0].resource_requirements[0].type #=> String, one of "GPU", "InferenceAccelerator", "NeuronDevice"
+    #   resp.task_definition.container_definitions[0].resource_requirements[0].sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.task_definition.container_definitions[0].resource_requirements[0].sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.task_definition.container_definitions[0].firelens_configuration.type #=> String, one of "fluentd", "fluentbit"
     #   resp.task_definition.container_definitions[0].firelens_configuration.options #=> Hash
     #   resp.task_definition.container_definitions[0].firelens_configuration.options["String"] #=> String
@@ -13051,8 +13087,14 @@ module Aws::ECS
     #           memory_reservation: 1,
     #           resource_requirements: [
     #             {
-    #               value: "String", # required
+    #               value: "String",
     #               type: "GPU", # required, accepts GPU, InferenceAccelerator, NeuronDevice
+    #               sharing_strategy: {
+    #                 nvidia_mps: {
+    #                   memory: 1, # required
+    #                   max_compute_percent: 1,
+    #                 },
+    #               },
     #             },
     #           ],
     #         },
@@ -13184,6 +13226,8 @@ module Aws::ECS
     #   resp.tasks[0].containers[0].gpu_ids[0] #=> String
     #   resp.tasks[0].containers[0].neuron_device_ids #=> Array
     #   resp.tasks[0].containers[0].neuron_device_ids[0] #=> String
+    #   resp.tasks[0].containers[0].gpu_sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.tasks[0].containers[0].gpu_sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.tasks[0].cpu #=> String
     #   resp.tasks[0].created_at #=> Time
     #   resp.tasks[0].desired_status #=> String
@@ -13213,6 +13257,8 @@ module Aws::ECS
     #   resp.tasks[0].overrides.container_overrides[0].resource_requirements #=> Array
     #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].value #=> String
     #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].type #=> String, one of "GPU", "InferenceAccelerator", "NeuronDevice"
+    #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.tasks[0].overrides.cpu #=> String
     #   resp.tasks[0].overrides.inference_accelerator_overrides #=> Array
     #   resp.tasks[0].overrides.inference_accelerator_overrides[0].device_name #=> String
@@ -13500,8 +13546,14 @@ module Aws::ECS
     #           memory_reservation: 1,
     #           resource_requirements: [
     #             {
-    #               value: "String", # required
+    #               value: "String",
     #               type: "GPU", # required, accepts GPU, InferenceAccelerator, NeuronDevice
+    #               sharing_strategy: {
+    #                 nvidia_mps: {
+    #                   memory: 1, # required
+    #                   max_compute_percent: 1,
+    #                 },
+    #               },
     #             },
     #           ],
     #         },
@@ -13619,6 +13671,8 @@ module Aws::ECS
     #   resp.tasks[0].containers[0].gpu_ids[0] #=> String
     #   resp.tasks[0].containers[0].neuron_device_ids #=> Array
     #   resp.tasks[0].containers[0].neuron_device_ids[0] #=> String
+    #   resp.tasks[0].containers[0].gpu_sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.tasks[0].containers[0].gpu_sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.tasks[0].cpu #=> String
     #   resp.tasks[0].created_at #=> Time
     #   resp.tasks[0].desired_status #=> String
@@ -13648,6 +13702,8 @@ module Aws::ECS
     #   resp.tasks[0].overrides.container_overrides[0].resource_requirements #=> Array
     #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].value #=> String
     #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].type #=> String, one of "GPU", "InferenceAccelerator", "NeuronDevice"
+    #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.tasks[0].overrides.container_overrides[0].resource_requirements[0].sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.tasks[0].overrides.cpu #=> String
     #   resp.tasks[0].overrides.inference_accelerator_overrides #=> Array
     #   resp.tasks[0].overrides.inference_accelerator_overrides[0].device_name #=> String
@@ -13928,6 +13984,8 @@ module Aws::ECS
     #   resp.task.containers[0].gpu_ids[0] #=> String
     #   resp.task.containers[0].neuron_device_ids #=> Array
     #   resp.task.containers[0].neuron_device_ids[0] #=> String
+    #   resp.task.containers[0].gpu_sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.task.containers[0].gpu_sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.task.cpu #=> String
     #   resp.task.created_at #=> Time
     #   resp.task.desired_status #=> String
@@ -13957,6 +14015,8 @@ module Aws::ECS
     #   resp.task.overrides.container_overrides[0].resource_requirements #=> Array
     #   resp.task.overrides.container_overrides[0].resource_requirements[0].value #=> String
     #   resp.task.overrides.container_overrides[0].resource_requirements[0].type #=> String, one of "GPU", "InferenceAccelerator", "NeuronDevice"
+    #   resp.task.overrides.container_overrides[0].resource_requirements[0].sharing_strategy.nvidia_mps.memory #=> Integer
+    #   resp.task.overrides.container_overrides[0].resource_requirements[0].sharing_strategy.nvidia_mps.max_compute_percent #=> Integer
     #   resp.task.overrides.cpu #=> String
     #   resp.task.overrides.inference_accelerator_overrides #=> Array
     #   resp.task.overrides.inference_accelerator_overrides[0].device_name #=> String
@@ -15185,6 +15245,12 @@ module Aws::ECS
     #   resp.container_instance.health_status.details[0].status_reason #=> String
     #   resp.container_instance.health_status.details[0].last_updated #=> Time
     #   resp.container_instance.health_status.details[0].last_status_change #=> Time
+    #   resp.container_instance.gpu_devices #=> Array
+    #   resp.container_instance.gpu_devices[0].gpu_id #=> String
+    #   resp.container_instance.gpu_devices[0].sharing_mode #=> String, one of "IDLE", "NVIDIA_MPS", "DEDICATED"
+    #   resp.container_instance.gpu_devices[0].total_memory_in_mi_b #=> Integer
+    #   resp.container_instance.gpu_devices[0].remaining_memory_in_mi_b #=> Integer
+    #   resp.container_instance.gpu_devices[0].container_count #=> Integer
     #
     # @see http://docs.aws.amazon.com/goto/WebAPI/ecs-2014-11-13/UpdateContainerAgent AWS API Documentation
     #
@@ -15513,6 +15579,12 @@ module Aws::ECS
     #   resp.container_instances[0].health_status.details[0].status_reason #=> String
     #   resp.container_instances[0].health_status.details[0].last_updated #=> Time
     #   resp.container_instances[0].health_status.details[0].last_status_change #=> Time
+    #   resp.container_instances[0].gpu_devices #=> Array
+    #   resp.container_instances[0].gpu_devices[0].gpu_id #=> String
+    #   resp.container_instances[0].gpu_devices[0].sharing_mode #=> String, one of "IDLE", "NVIDIA_MPS", "DEDICATED"
+    #   resp.container_instances[0].gpu_devices[0].total_memory_in_mi_b #=> Integer
+    #   resp.container_instances[0].gpu_devices[0].remaining_memory_in_mi_b #=> Integer
+    #   resp.container_instances[0].gpu_devices[0].container_count #=> Integer
     #   resp.failures #=> Array
     #   resp.failures[0].arn #=> String
     #   resp.failures[0].reason #=> String
@@ -17482,7 +17554,7 @@ module Aws::ECS
         tracer: tracer
       )
       context[:gem_name] = 'aws-sdk-ecs'
-      context[:gem_version] = '1.251.0'
+      context[:gem_version] = '1.252.0'
       Seahorse::Client::Request.new(handlers, context)
     end
 

data/lib/aws-sdk-ecs/client_api.rb

@@ -370,7 +370,10 @@ module Aws::ECS
     GatewayTestTrafficHeader = Shapes::StructureShape.new(name: 'GatewayTestTrafficHeader')
     GetTaskProtectionRequest = Shapes::StructureShape.new(name: 'GetTaskProtectionRequest')
     GetTaskProtectionResponse = Shapes::StructureShape.new(name: 'GetTaskProtectionResponse')
+    GpuDevice = Shapes::StructureShape.new(name: 'GpuDevice')
+    GpuDevices = Shapes::ListShape.new(name: 'GpuDevices')
     GpuIds = Shapes::ListShape.new(name: 'GpuIds')
+    GpuSharingMode = Shapes::StringShape.new(name: 'GpuSharingMode')
     HealthCheck = Shapes::StructureShape.new(name: 'HealthCheck')
     HealthStatus = Shapes::StringShape.new(name: 'HealthStatus')
     HealthyPercentInteger = Shapes::IntegerShape.new(name: 'HealthyPercentInteger')
@@ -508,6 +511,9 @@ module Aws::ECS
     NetworkMode = Shapes::StringShape.new(name: 'NetworkMode')
     NeuronDeviceIds = Shapes::ListShape.new(name: 'NeuronDeviceIds')
     NoUpdateAvailableException = Shapes::StructureShape.new(name: 'NoUpdateAvailableException')
+    NvidiaMpsAllocation = Shapes::StructureShape.new(name: 'NvidiaMpsAllocation')
+    NvidiaMpsAllocationMaxComputePercentInteger = Shapes::IntegerShape.new(name: 'NvidiaMpsAllocationMaxComputePercentInteger')
+    NvidiaMpsAllocationMemoryInteger = Shapes::IntegerShape.new(name: 'NvidiaMpsAllocationMemoryInteger')
     OSFamily = Shapes::StringShape.new(name: 'OSFamily')
     PidMode = Shapes::StringShape.new(name: 'PidMode')
     PlacementConstraint = Shapes::StructureShape.new(name: 'PlacementConstraint')
@@ -631,6 +637,7 @@ module Aws::ECS
     SettingName = Shapes::StringShape.new(name: 'SettingName')
     SettingType = Shapes::StringShape.new(name: 'SettingType')
     Settings = Shapes::ListShape.new(name: 'Settings')
+    SharingStrategy = Shapes::UnionShape.new(name: 'SharingStrategy')
     SortOrder = Shapes::StringShape.new(name: 'SortOrder')
     StabilityStatus = Shapes::StringShape.new(name: 'StabilityStatus')
     StartTaskRequest = Shapes::StructureShape.new(name: 'StartTaskRequest')
@@ -944,6 +951,7 @@ module Aws::ECS
     Container.add_member(:memory_reservation, Shapes::ShapeRef.new(shape: String, location_name: "memoryReservation"))
     Container.add_member(:gpu_ids, Shapes::ShapeRef.new(shape: GpuIds, location_name: "gpuIds"))
     Container.add_member(:neuron_device_ids, Shapes::ShapeRef.new(shape: NeuronDeviceIds, location_name: "neuronDeviceIds"))
+    Container.add_member(:gpu_sharing_strategy, Shapes::ShapeRef.new(shape: SharingStrategy, location_name: "gpuSharingStrategy"))
     Container.struct_class = Types::Container
 
     ContainerDefinition.add_member(:name, Shapes::ShapeRef.new(shape: String, location_name: "name"))
@@ -1023,6 +1031,7 @@ module Aws::ECS
     ContainerInstance.add_member(:attachments, Shapes::ShapeRef.new(shape: Attachments, location_name: "attachments"))
     ContainerInstance.add_member(:tags, Shapes::ShapeRef.new(shape: Tags, location_name: "tags"))
     ContainerInstance.add_member(:health_status, Shapes::ShapeRef.new(shape: ContainerInstanceHealthStatus, location_name: "healthStatus"))
+    ContainerInstance.add_member(:gpu_devices, Shapes::ShapeRef.new(shape: GpuDevices, location_name: "gpuDevices"))
     ContainerInstance.struct_class = Types::ContainerInstance
 
     ContainerInstanceFieldList.member = Shapes::ShapeRef.new(shape: ContainerInstanceField)
@@ -2136,6 +2145,15 @@ module Aws::ECS
     GetTaskProtectionResponse.add_member(:failures, Shapes::ShapeRef.new(shape: Failures, location_name: "failures"))
     GetTaskProtectionResponse.struct_class = Types::GetTaskProtectionResponse
 
+    GpuDevice.add_member(:gpu_id, Shapes::ShapeRef.new(shape: String, location_name: "gpuId"))
+    GpuDevice.add_member(:sharing_mode, Shapes::ShapeRef.new(shape: GpuSharingMode, location_name: "sharingMode"))
+    GpuDevice.add_member(:total_memory_in_mi_b, Shapes::ShapeRef.new(shape: BoxedInteger, location_name: "totalMemoryInMiB"))
+    GpuDevice.add_member(:remaining_memory_in_mi_b, Shapes::ShapeRef.new(shape: BoxedInteger, location_name: "remainingMemoryInMiB"))
+    GpuDevice.add_member(:container_count, Shapes::ShapeRef.new(shape: BoxedInteger, location_name: "containerCount"))
+    GpuDevice.struct_class = Types::GpuDevice
+
+    GpuDevices.member = Shapes::ShapeRef.new(shape: GpuDevice)
+
     GpuIds.member = Shapes::ShapeRef.new(shape: String)
 
     HealthCheck.add_member(:command, Shapes::ShapeRef.new(shape: StringList, required: true, location_name: "command"))
@@ -2688,6 +2706,10 @@ module Aws::ECS
     NoUpdateAvailableException.add_member(:message, Shapes::ShapeRef.new(shape: String, location_name: "message"))
     NoUpdateAvailableException.struct_class = Types::NoUpdateAvailableException
 
+    NvidiaMpsAllocation.add_member(:memory, Shapes::ShapeRef.new(shape: NvidiaMpsAllocationMemoryInteger, required: true, location_name: "memory"))
+    NvidiaMpsAllocation.add_member(:max_compute_percent, Shapes::ShapeRef.new(shape: NvidiaMpsAllocationMaxComputePercentInteger, location_name: "maxComputePercent"))
+    NvidiaMpsAllocation.struct_class = Types::NvidiaMpsAllocation
+
     PlacementConstraint.add_member(:type, Shapes::ShapeRef.new(shape: PlacementConstraintType, location_name: "type"))
     PlacementConstraint.add_member(:expression, Shapes::ShapeRef.new(shape: String, location_name: "expression"))
     PlacementConstraint.struct_class = Types::PlacementConstraint
@@ -2844,8 +2866,9 @@ module Aws::ECS
     ResourceNotFoundException.add_member(:message, Shapes::ShapeRef.new(shape: String, location_name: "message"))
     ResourceNotFoundException.struct_class = Types::ResourceNotFoundException
 
-    ResourceRequirement.add_member(:value, Shapes::ShapeRef.new(shape: String, required: true, location_name: "value"))
+    ResourceRequirement.add_member(:value, Shapes::ShapeRef.new(shape: String, location_name: "value"))
     ResourceRequirement.add_member(:type, Shapes::ShapeRef.new(shape: ResourceType, required: true, location_name: "type"))
+    ResourceRequirement.add_member(:sharing_strategy, Shapes::ShapeRef.new(shape: SharingStrategy, location_name: "sharingStrategy"))
     ResourceRequirement.struct_class = Types::ResourceRequirement
 
     ResourceRequirements.member = Shapes::ShapeRef.new(shape: ResourceRequirement)
@@ -3173,6 +3196,12 @@ module Aws::ECS
 
     Settings.member = Shapes::ShapeRef.new(shape: Setting)
 
+    SharingStrategy.add_member(:nvidia_mps, Shapes::ShapeRef.new(shape: NvidiaMpsAllocation, location_name: "nvidiaMps"))
+    SharingStrategy.add_member(:unknown, Shapes::ShapeRef.new(shape: nil, location_name: 'unknown'))
+    SharingStrategy.add_member_subclass(:nvidia_mps, Types::SharingStrategy::NvidiaMps)
+    SharingStrategy.add_member_subclass(:unknown, Types::SharingStrategy::Unknown)
+    SharingStrategy.struct_class = Types::SharingStrategy
+
     StartTaskRequest.add_member(:cluster, Shapes::ShapeRef.new(shape: String, location_name: "cluster"))
     StartTaskRequest.add_member(:container_instances, Shapes::ShapeRef.new(shape: StringList, required: true, location_name: "containerInstances"))
     StartTaskRequest.add_member(:enable_ecs_managed_tags, Shapes::ShapeRef.new(shape: Boolean, location_name: "enableECSManagedTags"))

data/lib/aws-sdk-ecs/types.rb

@@ -1394,6 +1394,12 @@ module Aws::ECS
     #   The IDs of each Neuron device assigned to the container.
     #   @return [Array<String>]
     #
+    # @!attribute [rw] gpu_sharing_strategy
+    #   The GPU sharing strategy that the task definition specifies for the
+    #   container. This field is present only for containers whose resource
+    #   requirements include `sharingStrategy`.
+    #   @return [Types::SharingStrategy]
+    #
     # @see http://docs.aws.amazon.com/goto/WebAPI/ecs-2014-11-13/Container AWS API Documentation
     #
     class Container < Struct.new(
@@ -1414,7 +1420,8 @@ module Aws::ECS
       :memory,
       :memory_reservation,
       :gpu_ids,
-      :neuron_device_ids)
+      :neuron_device_ids,
+      :gpu_sharing_strategy)
       SENSITIVE = []
       include Aws::Structure
     end
@@ -2567,6 +2574,12 @@ module Aws::ECS
     #   An object representing the health status of the container instance.
     #   @return [Types::ContainerInstanceHealthStatus]
     #
+    # @!attribute [rw] gpu_devices
+    #   The state of each GPU on the container instance, including its
+    #   sharing mode and available memory. This field is present only for
+    #   container instances that support GPU sharing.
+    #   @return [Array<Types::GpuDevice>]
+    #
     # @see http://docs.aws.amazon.com/goto/WebAPI/ecs-2014-11-13/ContainerInstance AWS API Documentation
     #
     class ContainerInstance < Struct.new(
@@ -2587,7 +2600,8 @@ module Aws::ECS
       :registered_at,
       :attachments,
       :tags,
-      :health_status)
+      :health_status,
+      :gpu_devices)
       SENSITIVE = []
       include Aws::Structure
     end
@@ -9607,6 +9621,61 @@ module Aws::ECS
       include Aws::Structure
     end
 
+    # The state of a GPU on a container instance, including its sharing mode
+    # and the memory that's available for new containers.
+    #
+    # @!attribute [rw] gpu_id
+    #   The ID of the GPU. This ID matches the GPU IDs in the `gpuIds` of
+    #   the containers that Amazon ECS assigned to the GPU.
+    #   @return [String]
+    #
+    # @!attribute [rw] sharing_mode
+    #   How the GPU is shared between containers.
+    #
+    #   Valid values:
+    #
+    #   * `IDLE` - No containers are assigned to the GPU, and all of its
+    #     memory is available.
+    #
+    #   * `NVIDIA_MPS` - One or more containers share the GPU by using
+    #     NVIDIA Multi-Process Service (MPS).
+    #
+    #   * `DEDICATED` - One container has exclusive use of the GPU.
+    #   @return [String]
+    #
+    # @!attribute [rw] total_memory_in_mi_b
+    #   The total amount of GPU memory, in MiB, that the Amazon ECS
+    #   container agent reported for the GPU. This is the memory that's
+    #   available to containers, and it can be less than the nominal memory
+    #   of the GPU.
+    #   @return [Integer]
+    #
+    # @!attribute [rw] remaining_memory_in_mi_b
+    #   The amount of GPU memory, in MiB, that's available to allocate to
+    #   new containers that use the `nvidiaMps` sharing strategy. Amazon ECS
+    #   calculates this value from the memory that it allocated to
+    #   containers on the GPU, not from the memory that the containers use.
+    #   @return [Integer]
+    #
+    # @!attribute [rw] container_count
+    #   The number of containers that Amazon ECS assigned to the GPU. This
+    #   value is `0` when the sharing mode is `IDLE` and `1` when the
+    #   sharing mode is `DEDICATED`. Amazon ECS counts the containers that
+    #   it assigned to the GPU, not the processes that are running on it.
+    #   @return [Integer]
+    #
+    # @see http://docs.aws.amazon.com/goto/WebAPI/ecs-2014-11-13/GpuDevice AWS API Documentation
+    #
+    class GpuDevice < Struct.new(
+      :gpu_id,
+      :sharing_mode,
+      :total_memory_in_mi_b,
+      :remaining_memory_in_mi_b,
+      :container_count)
+      SENSITIVE = []
+      include Aws::Structure
+    end
+
     # An object representing a container health check. Health check
     # parameters that are specified in a container definition override any
     # Docker health checks that exist in the container image (such as those
@@ -13685,6 +13754,43 @@ module Aws::ECS
       include Aws::Structure
     end
 
+    # The GPU memory and compute limits for a container that shares a GPU by
+    # using NVIDIA Multi-Process Service (MPS). Amazon ECS assigns the
+    # container a single GPU that has enough memory available. Other
+    # containers that use MPS can share the same GPU.
+    #
+    # @!attribute [rw] memory
+    #   The amount of GPU memory, in MiB, to allocate to the container.
+    #   Amazon ECS assigns the container a GPU that has at least the
+    #   specified amount of memory available, and sets the
+    #   `CUDA_MPS_PINNED_DEVICE_MEM_LIMIT` environment variable in the
+    #   container to limit its memory to this amount. The memory must be
+    #   available on a single GPU. If more than one container in a task uses
+    #   `nvidiaMps`, Amazon ECS assigns them the same GPU, which must have
+    #   enough memory available for all of them.
+    #   @return [Integer]
+    #
+    # @!attribute [rw] max_compute_percent
+    #   The maximum percentage of the GPU's compute capacity that the
+    #   container can use. Amazon ECS sets the
+    #   `CUDA_MPS_ACTIVE_THREAD_PERCENTAGE` environment variable in the
+    #   container to this value. This value is a limit, not a reservation.
+    #
+    #   If you don't specify this parameter, Amazon ECS doesn't set
+    #   `CUDA_MPS_ACTIVE_THREAD_PERCENTAGE`, and the active thread
+    #   percentage of the MPS server on the container instance applies. By
+    #   default, this is 100 percent.
+    #   @return [Integer]
+    #
+    # @see http://docs.aws.amazon.com/goto/WebAPI/ecs-2014-11-13/NvidiaMpsAllocation AWS API Documentation
+    #
+    class NvidiaMpsAllocation < Struct.new(
+      :memory,
+      :max_compute_percent)
+      SENSITIVE = []
+      include Aws::Structure
+    end
+
     # An object representing a constraint on task placement. For more
     # information, see [Task placement constraints][1] in the *Amazon
     # Elastic Container Service Developer Guide*.
@@ -15366,6 +15472,9 @@ module Aws::ECS
     # @!attribute [rw] value
     #   The value for the specified resource type.
     #
+    #   You must specify this parameter unless the type is `GPU` and you
+    #   specify `sharingStrategy`.
+    #
     #   When the type is `GPU`, the value is the number of physical `GPUs`
     #   the Amazon ECS container agent reserves for the container. The
     #   number of GPUs that's reserved for all containers in a task can't
@@ -15392,11 +15501,26 @@ module Aws::ECS
     #   The type of resource to assign to a container.
     #   @return [String]
     #
+    # @!attribute [rw] sharing_strategy
+    #   The strategy that the container uses to share a GPU with other
+    #   containers. Specify a sharing strategy only when the type is `GPU`.
+    #   For a GPU resource requirement, specify either `value` or
+    #   `sharingStrategy`, but not both. If you don't specify a sharing
+    #   strategy, the container has exclusive use of each GPU that Amazon
+    #   ECS assigns to it. The containers in a task that request GPUs must
+    #   all use the same option.
+    #
+    #   You can use GPU sharing with tasks that run on Amazon EC2 instances
+    #   or Amazon ECS Managed Instances. Task definitions that require
+    #   `EXTERNAL` compatibility don't support GPU sharing.
+    #   @return [Types::SharingStrategy]
+    #
     # @see http://docs.aws.amazon.com/goto/WebAPI/ecs-2014-11-13/ResourceRequirement AWS API Documentation
     #
     class ResourceRequirement < Struct.new(
       :value,
-      :type)
+      :type,
+      :sharing_strategy)
       SENSITIVE = []
       include Aws::Structure
     end
@@ -17919,6 +18043,33 @@ module Aws::ECS
       include Aws::Structure
     end
 
+    # The strategy that a container uses to share a GPU with other
+    # containers. Specify exactly one member.
+    #
+    # @note SharingStrategy is a union - when making an API calls you must set exactly one of the members.
+    #
+    # @note SharingStrategy is a union - when returned from an API call exactly one value will be set and the returned type will be a subclass of SharingStrategy corresponding to the set member.
+    #
+    # @!attribute [rw] nvidia_mps
+    #   Shares a GPU between containers by using NVIDIA Multi-Process
+    #   Service (MPS). You specify the GPU memory to allocate to the
+    #   container and, optionally, the maximum percentage of the GPU's
+    #   compute capacity that the container can use.
+    #   @return [Types::NvidiaMpsAllocation]
+    #
+    # @see http://docs.aws.amazon.com/goto/WebAPI/ecs-2014-11-13/SharingStrategy AWS API Documentation
+    #
+    class SharingStrategy < Struct.new(
+      :nvidia_mps,
+      :unknown)
+      SENSITIVE = []
+      include Aws::Structure
+      include Aws::Structure::Union
+
+      class NvidiaMps < SharingStrategy; end
+      class Unknown < SharingStrategy; end
+    end
+
     # <zonbook />
     #
     # <xhtml />

data/lib/aws-sdk-ecs.rb

@@ -55,7 +55,7 @@ module Aws::ECS
   autoload :EndpointProvider, 'aws-sdk-ecs/endpoint_provider'
   autoload :Endpoints, 'aws-sdk-ecs/endpoints'
 
-  GEM_VERSION = '1.251.0'
+  GEM_VERSION = '1.252.0'
 
 end
 

data/sig/client.rbs

@@ -1506,10 +1506,7 @@ module Aws
                                             }
                                           ]?,
                                           resource_requirements: Array[
-                                            {
-                                              value: ::String,
-                                              type: ("GPU" | "InferenceAccelerator" | "NeuronDevice")
-                                            }
+                                            Params::resource_requirement
                                           ]?,
                                           firelens_configuration: {
                                             type: ("fluentd" | "fluentbit"),

data/sig/params.rbs

@@ -319,6 +319,17 @@ module Aws
       }?
     }
 
+      type resource_requirement = {
+      value: ::String?,
+      type: ("GPU" | "InferenceAccelerator" | "NeuronDevice"),
+      sharing_strategy: {
+        nvidia_mps: {
+          memory: ::Integer,
+          max_compute_percent: ::Integer?
+        }?
+      }?
+    }
+
       type task_override = {
       container_overrides: Array[
         Params::container_override
@@ -357,10 +368,7 @@ module Aws
       memory: ::Integer?,
       memory_reservation: ::Integer?,
       resource_requirements: Array[
-        {
-          value: ::String,
-          type: ("GPU" | "InferenceAccelerator" | "NeuronDevice")
-        }
+        Params::resource_requirement
       ]?
     }
 

data/sig/types.rbs

@@ -241,6 +241,7 @@ module Aws::ECS
       attr_accessor memory_reservation: ::String
       attr_accessor gpu_ids: ::Array[::String]
       attr_accessor neuron_device_ids: ::Array[::String]
+      attr_accessor gpu_sharing_strategy: Types::SharingStrategy
       SENSITIVE: []
     end
 
@@ -322,6 +323,7 @@ module Aws::ECS
       attr_accessor attachments: ::Array[Types::Attachment]
       attr_accessor tags: ::Array[Types::Tag]
       attr_accessor health_status: Types::ContainerInstanceHealthStatus
+      attr_accessor gpu_devices: ::Array[Types::GpuDevice]
       SENSITIVE: []
     end
 
@@ -1704,6 +1706,15 @@ module Aws::ECS
       SENSITIVE: []
     end
 
+    class GpuDevice
+      attr_accessor gpu_id: ::String
+      attr_accessor sharing_mode: ("IDLE" | "NVIDIA_MPS" | "DEDICATED")
+      attr_accessor total_memory_in_mi_b: ::Integer
+      attr_accessor remaining_memory_in_mi_b: ::Integer
+      attr_accessor container_count: ::Integer
+      SENSITIVE: []
+    end
+
     class HealthCheck
       attr_accessor command: ::Array[::String]
       attr_accessor interval: ::Integer
@@ -2379,6 +2390,12 @@ module Aws::ECS
       SENSITIVE: []
     end
 
+    class NvidiaMpsAllocation
+      attr_accessor memory: ::Integer
+      attr_accessor max_compute_percent: ::Integer
+      SENSITIVE: []
+    end
+
     class PlacementConstraint
       attr_accessor type: ("distinctInstance" | "memberOf")
       attr_accessor expression: ::String
@@ -2576,6 +2593,7 @@ module Aws::ECS
     class ResourceRequirement
       attr_accessor value: ::String
       attr_accessor type: ("GPU" | "InferenceAccelerator" | "NeuronDevice")
+      attr_accessor sharing_strategy: Types::SharingStrategy
       SENSITIVE: []
     end
 
@@ -2945,6 +2963,17 @@ module Aws::ECS
       SENSITIVE: []
     end
 
+    class SharingStrategy
+      attr_accessor nvidia_mps: Types::NvidiaMpsAllocation
+      attr_accessor unknown: untyped
+      SENSITIVE: []
+
+      class NvidiaMps < SharingStrategy
+      end
+      class Unknown < SharingStrategy
+      end
+    end
+
     class StartTaskRequest
       attr_accessor cluster: ::String
       attr_accessor container_instances: ::Array[::String]

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: aws-sdk-ecs
 version: !ruby/object:Gem::Version
-  version: 1.251.0
+  version: 1.252.0
 platform: ruby
 authors:
 - Amazon Web Services