aws-sdk-sagemaker 1.383.0 → 1.384.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 3e818bd870be19b6b855c0d751115d66ff42249f2c7fb2d150459f8c1c62cc8c
-  data.tar.gz: 5dfb5180298731ccf458258afabf6e1c6d233d6d14e00983291b6ebbb8bd598c
+  metadata.gz: a3266009736b6b5612256074edfbd19cf4b6e9c2c8787b25a4326464171ede3a
+  data.tar.gz: 0c63a8ed4cae74a43f4822b4d9df2628ee207dc99d17cbaac0f620c4d9c88840
 SHA512:
-  metadata.gz: 645f7c91235ef6f2cae7a286e2009c579d0bbee49452b7d457e76bcdda849e7181f9ec4a2a9fe0bb98bb43dfaca9f0a59f6149c02ce401e573bdccac9b5331e2
-  data.tar.gz: 92baeedd9c315d8c6604178588fd6c61d36fcbcd913ea6308c44954967f90770072e920f5a93aee557d7ce790ca84aa7e7fdebe5d7fe4ca2d6e69aec93605706
+  metadata.gz: 7b1c0c9510476ad1c4ef9b80f6541e31bba49e12ddddb6fa40887c36b9c71d33d5d25fc9302bafdd3c1d3a333efbd8075a6098b074d33ac400512912ecb97c08
+  data.tar.gz: 6697605752bb56f096498c98757184f0febed63850417133347d1a7fab528651231a71adc272a40812f92020863722d2a53eb1c0c72a90ce1d728f0261d759eb

data/CHANGELOG.md

@@ -1,6 +1,11 @@
 Unreleased Changes
 ------------------
 
+1.384.0 (2026-08-10)
+------------------
+
+* Feature - Added PREFIX AWARE routing strategy and PrefixAwareRoutingConfig to CreateEndpointConfig. Configure PrefixLength and ConcurrencyThreshold to route requests that share the same prompt prefix to the same instance.
+
 1.383.0 (2026-08-07)
 ------------------
 

data/VERSION

@@ -1 +1 @@
-1.383.0
+1.384.0

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

@@ -4823,7 +4823,11 @@ module Aws::SageMaker
     #           },
     #         },
     #         routing_config: {
-    #           routing_strategy: "LEAST_OUTSTANDING_REQUESTS", # required, accepts LEAST_OUTSTANDING_REQUESTS, RANDOM
+    #           routing_strategy: "LEAST_OUTSTANDING_REQUESTS", # required, accepts LEAST_OUTSTANDING_REQUESTS, RANDOM, PREFIX_AWARE
+    #           prefix_aware_routing_config: {
+    #             prefix_length: 1,
+    #             concurrency_threshold: 1,
+    #           },
     #         },
     #         inference_ami_version: "al2-ami-sagemaker-inference-gpu-2", # accepts al2-ami-sagemaker-inference-gpu-2, al2-ami-sagemaker-inference-gpu-2-1, al2-ami-sagemaker-inference-gpu-3-1, al2-ami-sagemaker-inference-neuron-2, al2023-ami-sagemaker-inference-gpu-4-1
     #         capacity_reservation_config: {
@@ -4941,7 +4945,11 @@ module Aws::SageMaker
     #           },
     #         },
     #         routing_config: {
-    #           routing_strategy: "LEAST_OUTSTANDING_REQUESTS", # required, accepts LEAST_OUTSTANDING_REQUESTS, RANDOM
+    #           routing_strategy: "LEAST_OUTSTANDING_REQUESTS", # required, accepts LEAST_OUTSTANDING_REQUESTS, RANDOM, PREFIX_AWARE
+    #           prefix_aware_routing_config: {
+    #             prefix_length: 1,
+    #             concurrency_threshold: 1,
+    #           },
     #         },
     #         inference_ami_version: "al2-ami-sagemaker-inference-gpu-2", # accepts al2-ami-sagemaker-inference-gpu-2, al2-ami-sagemaker-inference-gpu-2-1, al2-ami-sagemaker-inference-gpu-3-1, al2-ami-sagemaker-inference-neuron-2, al2023-ami-sagemaker-inference-gpu-4-1
     #         capacity_reservation_config: {
@@ -16346,7 +16354,9 @@ module Aws::SageMaker
     #   resp.production_variants[0].managed_instance_scaling.scale_in_policy.strategy #=> String, one of "IDLE_RELEASE", "CONSOLIDATION"
     #   resp.production_variants[0].managed_instance_scaling.scale_in_policy.maximum_step_size #=> Integer
     #   resp.production_variants[0].managed_instance_scaling.scale_in_policy.cooldown_in_minutes #=> Integer
-    #   resp.production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM"
+    #   resp.production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM", "PREFIX_AWARE"
+    #   resp.production_variants[0].routing_config.prefix_aware_routing_config.prefix_length #=> Integer
+    #   resp.production_variants[0].routing_config.prefix_aware_routing_config.concurrency_threshold #=> Integer
     #   resp.production_variants[0].capacity_reservation_config.ml_reservation_arn #=> String
     #   resp.production_variants[0].capacity_reservation_config.capacity_reservation_preference #=> String, one of "capacity-reservations-only"
     #   resp.production_variants[0].capacity_reservation_config.total_instance_count #=> Integer
@@ -16422,7 +16432,9 @@ module Aws::SageMaker
     #   resp.pending_deployment_summary.production_variants[0].managed_instance_scaling.scale_in_policy.strategy #=> String, one of "IDLE_RELEASE", "CONSOLIDATION"
     #   resp.pending_deployment_summary.production_variants[0].managed_instance_scaling.scale_in_policy.maximum_step_size #=> Integer
     #   resp.pending_deployment_summary.production_variants[0].managed_instance_scaling.scale_in_policy.cooldown_in_minutes #=> Integer
-    #   resp.pending_deployment_summary.production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM"
+    #   resp.pending_deployment_summary.production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM", "PREFIX_AWARE"
+    #   resp.pending_deployment_summary.production_variants[0].routing_config.prefix_aware_routing_config.prefix_length #=> Integer
+    #   resp.pending_deployment_summary.production_variants[0].routing_config.prefix_aware_routing_config.concurrency_threshold #=> Integer
     #   resp.pending_deployment_summary.start_time #=> Time
     #   resp.pending_deployment_summary.shadow_production_variants #=> Array
     #   resp.pending_deployment_summary.shadow_production_variants[0].variant_name #=> String
@@ -16455,7 +16467,9 @@ module Aws::SageMaker
     #   resp.pending_deployment_summary.shadow_production_variants[0].managed_instance_scaling.scale_in_policy.strategy #=> String, one of "IDLE_RELEASE", "CONSOLIDATION"
     #   resp.pending_deployment_summary.shadow_production_variants[0].managed_instance_scaling.scale_in_policy.maximum_step_size #=> Integer
     #   resp.pending_deployment_summary.shadow_production_variants[0].managed_instance_scaling.scale_in_policy.cooldown_in_minutes #=> Integer
-    #   resp.pending_deployment_summary.shadow_production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM"
+    #   resp.pending_deployment_summary.shadow_production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM", "PREFIX_AWARE"
+    #   resp.pending_deployment_summary.shadow_production_variants[0].routing_config.prefix_aware_routing_config.prefix_length #=> Integer
+    #   resp.pending_deployment_summary.shadow_production_variants[0].routing_config.prefix_aware_routing_config.concurrency_threshold #=> Integer
     #   resp.explainer_config.clarify_explainer_config.enable_explanations #=> String
     #   resp.explainer_config.clarify_explainer_config.inference_config.features_attribute #=> String
     #   resp.explainer_config.clarify_explainer_config.inference_config.content_template #=> String
@@ -16508,7 +16522,9 @@ module Aws::SageMaker
     #   resp.shadow_production_variants[0].managed_instance_scaling.scale_in_policy.strategy #=> String, one of "IDLE_RELEASE", "CONSOLIDATION"
     #   resp.shadow_production_variants[0].managed_instance_scaling.scale_in_policy.maximum_step_size #=> Integer
     #   resp.shadow_production_variants[0].managed_instance_scaling.scale_in_policy.cooldown_in_minutes #=> Integer
-    #   resp.shadow_production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM"
+    #   resp.shadow_production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM", "PREFIX_AWARE"
+    #   resp.shadow_production_variants[0].routing_config.prefix_aware_routing_config.prefix_length #=> Integer
+    #   resp.shadow_production_variants[0].routing_config.prefix_aware_routing_config.concurrency_threshold #=> Integer
     #   resp.shadow_production_variants[0].capacity_reservation_config.ml_reservation_arn #=> String
     #   resp.shadow_production_variants[0].capacity_reservation_config.capacity_reservation_preference #=> String, one of "capacity-reservations-only"
     #   resp.shadow_production_variants[0].capacity_reservation_config.total_instance_count #=> Integer
@@ -16597,7 +16613,9 @@ module Aws::SageMaker
     #   resp.production_variants[0].managed_instance_scaling.scale_in_policy.strategy #=> String, one of "IDLE_RELEASE", "CONSOLIDATION"
     #   resp.production_variants[0].managed_instance_scaling.scale_in_policy.maximum_step_size #=> Integer
     #   resp.production_variants[0].managed_instance_scaling.scale_in_policy.cooldown_in_minutes #=> Integer
-    #   resp.production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM"
+    #   resp.production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM", "PREFIX_AWARE"
+    #   resp.production_variants[0].routing_config.prefix_aware_routing_config.prefix_length #=> Integer
+    #   resp.production_variants[0].routing_config.prefix_aware_routing_config.concurrency_threshold #=> Integer
     #   resp.production_variants[0].inference_ami_version #=> String, one of "al2-ami-sagemaker-inference-gpu-2", "al2-ami-sagemaker-inference-gpu-2-1", "al2-ami-sagemaker-inference-gpu-3-1", "al2-ami-sagemaker-inference-neuron-2", "al2023-ami-sagemaker-inference-gpu-4-1"
     #   resp.production_variants[0].capacity_reservation_config.capacity_reservation_preference #=> String, one of "capacity-reservations-only"
     #   resp.production_variants[0].capacity_reservation_config.ml_reservation_arn #=> String
@@ -16671,7 +16689,9 @@ module Aws::SageMaker
     #   resp.shadow_production_variants[0].managed_instance_scaling.scale_in_policy.strategy #=> String, one of "IDLE_RELEASE", "CONSOLIDATION"
     #   resp.shadow_production_variants[0].managed_instance_scaling.scale_in_policy.maximum_step_size #=> Integer
     #   resp.shadow_production_variants[0].managed_instance_scaling.scale_in_policy.cooldown_in_minutes #=> Integer
-    #   resp.shadow_production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM"
+    #   resp.shadow_production_variants[0].routing_config.routing_strategy #=> String, one of "LEAST_OUTSTANDING_REQUESTS", "RANDOM", "PREFIX_AWARE"
+    #   resp.shadow_production_variants[0].routing_config.prefix_aware_routing_config.prefix_length #=> Integer
+    #   resp.shadow_production_variants[0].routing_config.prefix_aware_routing_config.concurrency_threshold #=> Integer
     #   resp.shadow_production_variants[0].inference_ami_version #=> String, one of "al2-ami-sagemaker-inference-gpu-2", "al2-ami-sagemaker-inference-gpu-2-1", "al2-ami-sagemaker-inference-gpu-3-1", "al2-ami-sagemaker-inference-neuron-2", "al2023-ami-sagemaker-inference-gpu-4-1"
     #   resp.shadow_production_variants[0].capacity_reservation_config.capacity_reservation_preference #=> String, one of "capacity-reservations-only"
     #   resp.shadow_production_variants[0].capacity_reservation_config.ml_reservation_arn #=> String
@@ -34563,7 +34583,7 @@ module Aws::SageMaker
         tracer: tracer
       )
       context[:gem_name] = 'aws-sdk-sagemaker'
-      context[:gem_version] = '1.383.0'
+      context[:gem_version] = '1.384.0'
       Seahorse::Client::Request.new(handlers, context)
     end
 

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

@@ -2250,6 +2250,9 @@ module Aws::SageMaker
     PolicyString = Shapes::StringShape.new(name: 'PolicyString')
     PredefinedMetricSpecification = Shapes::StructureShape.new(name: 'PredefinedMetricSpecification')
     PreemptTeamTasks = Shapes::StringShape.new(name: 'PreemptTeamTasks')
+    PrefixAwareRoutingConcurrencyThreshold = Shapes::IntegerShape.new(name: 'PrefixAwareRoutingConcurrencyThreshold')
+    PrefixAwareRoutingConfig = Shapes::StructureShape.new(name: 'PrefixAwareRoutingConfig')
+    PrefixAwareRoutingPrefixLength = Shapes::IntegerShape.new(name: 'PrefixAwareRoutingPrefixLength')
     PresignedDomainUrl = Shapes::StringShape.new(name: 'PresignedDomainUrl')
     PresignedUrlAccessConfig = Shapes::StructureShape.new(name: 'PresignedUrlAccessConfig')
     PriorityClass = Shapes::StructureShape.new(name: 'PriorityClass')
@@ -11149,6 +11152,10 @@ module Aws::SageMaker
     PredefinedMetricSpecification.add_member(:predefined_metric_type, Shapes::ShapeRef.new(shape: String, location_name: "PredefinedMetricType"))
     PredefinedMetricSpecification.struct_class = Types::PredefinedMetricSpecification
 
+    PrefixAwareRoutingConfig.add_member(:prefix_length, Shapes::ShapeRef.new(shape: PrefixAwareRoutingPrefixLength, location_name: "PrefixLength"))
+    PrefixAwareRoutingConfig.add_member(:concurrency_threshold, Shapes::ShapeRef.new(shape: PrefixAwareRoutingConcurrencyThreshold, location_name: "ConcurrencyThreshold"))
+    PrefixAwareRoutingConfig.struct_class = Types::PrefixAwareRoutingConfig
+
     PresignedUrlAccessConfig.add_member(:accept_eula, Shapes::ShapeRef.new(shape: Boolean, location_name: "AcceptEula", metadata: {"box" => true}))
     PresignedUrlAccessConfig.add_member(:expected_s3_url, Shapes::ShapeRef.new(shape: S3ModelUri, location_name: "ExpectedS3Url"))
     PresignedUrlAccessConfig.struct_class = Types::PresignedUrlAccessConfig
@@ -11301,6 +11308,7 @@ module Aws::SageMaker
     ProductionVariantManagedInstanceScalingScaleInPolicy.struct_class = Types::ProductionVariantManagedInstanceScalingScaleInPolicy
 
     ProductionVariantRoutingConfig.add_member(:routing_strategy, Shapes::ShapeRef.new(shape: RoutingStrategy, required: true, location_name: "RoutingStrategy"))
+    ProductionVariantRoutingConfig.add_member(:prefix_aware_routing_config, Shapes::ShapeRef.new(shape: PrefixAwareRoutingConfig, location_name: "PrefixAwareRoutingConfig"))
     ProductionVariantRoutingConfig.struct_class = Types::ProductionVariantRoutingConfig
 
     ProductionVariantServerlessConfig.add_member(:memory_size_in_mb, Shapes::ShapeRef.new(shape: ServerlessMemorySizeInMB, required: true, location_name: "MemorySizeInMB"))

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

@@ -46189,6 +46189,44 @@ module Aws::SageMaker
       include Aws::Structure
     end
 
+    # The configuration for prefix-aware routing on a SageMaker real-time
+    # inference endpoint. Specify `PrefixLength` and `ConcurrencyThreshold`
+    # to control routing behavior.
+    #
+    # @!attribute [rw] prefix_length
+    #   The maximum length of the prefix used for routing decisions.
+    #   Required when `RoutingStrategy` is `PREFIX_AWARE`.
+    #
+    #   * For the SageMaker Runtime `InvokeEndpoint` and
+    #     `InvokeEndpointWithResponseStream` APIs, this value specifies the
+    #     number of bytes from the beginning of the request body.
+    #
+    #   * For OpenAI-compatible API, this value specifies the number of
+    #     characters from the text content of the messages array.
+    #
+    #   The endpoint routes requests that share the same prefix to the same
+    #   instance. Set this value to cover shared content (such as system
+    #   prompts) plus enough unique content to distribute workloads across
+    #   instances.
+    #   @return [Integer]
+    #
+    # @!attribute [rw] concurrency_threshold
+    #   The maximum number of in-flight requests on the target instance
+    #   before the endpoint routes to another instance. Required when
+    #   `RoutingStrategy` is `PREFIX_AWARE`. When in-flight requests on the
+    #   prefix-selected instance reach this threshold, the endpoint routes
+    #   the request to an instance with more available capacity.
+    #   @return [Integer]
+    #
+    # @see http://docs.aws.amazon.com/goto/WebAPI/sagemaker-2017-07-24/PrefixAwareRoutingConfig AWS API Documentation
+    #
+    class PrefixAwareRoutingConfig < Struct.new(
+      :prefix_length,
+      :concurrency_threshold)
+      SENSITIVE = []
+      include Aws::Structure
+    end
+
     # Configuration for accessing hub content through presigned URLs,
     # including license agreement acceptance and URL validation settings.
     #
@@ -47189,12 +47227,24 @@ module Aws::SageMaker
     #
     #   * `RANDOM`: The endpoint routes each request to a randomly chosen
     #     instance.
+    #
+    #   * `PREFIX_AWARE`: The endpoint routes requests that share the same
+    #     prompt prefix to the same instance. When the number of in-flight
+    #     requests on the selected instance reaches the configured
+    #     threshold, the endpoint routes the request to an instance with
+    #     more available capacity.
     #   @return [String]
     #
+    # @!attribute [rw] prefix_aware_routing_config
+    #   The configuration for prefix-aware routing. Specify this parameter
+    #   only when you set `RoutingStrategy` to `PREFIX_AWARE`.
+    #   @return [Types::PrefixAwareRoutingConfig]
+    #
     # @see http://docs.aws.amazon.com/goto/WebAPI/sagemaker-2017-07-24/ProductionVariantRoutingConfig AWS API Documentation
     #
     class ProductionVariantRoutingConfig < Struct.new(
-      :routing_strategy)
+      :routing_strategy,
+      :prefix_aware_routing_config)
       SENSITIVE = []
       include Aws::Structure
     end

data/lib/aws-sdk-sagemaker.rb

@@ -55,7 +55,7 @@ module Aws::SageMaker
   autoload :EndpointProvider, 'aws-sdk-sagemaker/endpoint_provider'
   autoload :Endpoints, 'aws-sdk-sagemaker/endpoints'
 
-  GEM_VERSION = '1.383.0'
+  GEM_VERSION = '1.384.0'
 
 end
 

data/sig/params.rbs

@@ -801,7 +801,11 @@ module Aws
       enable_ssm_access: bool?,
       managed_instance_scaling: Params::production_variant_managed_instance_scaling?,
       routing_config: {
-        routing_strategy: ("LEAST_OUTSTANDING_REQUESTS" | "RANDOM")
+        routing_strategy: ("LEAST_OUTSTANDING_REQUESTS" | "RANDOM" | "PREFIX_AWARE"),
+        prefix_aware_routing_config: {
+          prefix_length: ::Integer?,
+          concurrency_threshold: ::Integer?
+        }?
       }?,
       inference_ami_version: ("al2-ami-sagemaker-inference-gpu-2" | "al2-ami-sagemaker-inference-gpu-2-1" | "al2-ami-sagemaker-inference-gpu-3-1" | "al2-ami-sagemaker-inference-neuron-2" | "al2023-ami-sagemaker-inference-gpu-4-1")?,
       capacity_reservation_config: {

data/sig/types.rbs

@@ -9937,6 +9937,12 @@ module Aws::SageMaker
       SENSITIVE: []
     end
 
+    class PrefixAwareRoutingConfig
+      attr_accessor prefix_length: ::Integer
+      attr_accessor concurrency_threshold: ::Integer
+      SENSITIVE: []
+    end
+
     class PresignedUrlAccessConfig
       attr_accessor accept_eula: bool
       attr_accessor expected_s3_url: ::String
@@ -10114,7 +10120,8 @@ module Aws::SageMaker
     end
 
     class ProductionVariantRoutingConfig
-      attr_accessor routing_strategy: ("LEAST_OUTSTANDING_REQUESTS" | "RANDOM")
+      attr_accessor routing_strategy: ("LEAST_OUTSTANDING_REQUESTS" | "RANDOM" | "PREFIX_AWARE")
+      attr_accessor prefix_aware_routing_config: Types::PrefixAwareRoutingConfig
       SENSITIVE: []
     end
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: aws-sdk-sagemaker
 version: !ruby/object:Gem::Version
-  version: 1.383.0
+  version: 1.384.0
 platform: ruby
 authors:
 - Amazon Web Services