Create an Amazon SageMaker inference endpoint

Auth

Create an inference endpoint to perform an inference task with the amazon_sagemaker service.

Required authorization

Cluster privileges manage_inference

Path parameters

task_type enum required

The type of the inference task that the model will perform.

Values: text_embedding completion chat_completion sparse_embedding rerank
amazonsagemaker_inference_id string required

The unique identifier of the inference endpoint.

Query parameters

timeout string · default: 30s

Specifies the amount of time to wait for the inference endpoint to be created.

Request body

chunking_settings object

The chunking configuration object. Applies only to the sparse_embedding or text_embedding task types. Not applicable to the rerank, completion, or chat_completion task types.

Reference
max_chunk_size number · default: 250

The maximum size of a chunk in words. This value cannot be lower than 20 (for sentence strategy) or 10 (for word strategy). This value should not exceed the window size for the associated model.

overlap number · default: 100

The number of overlapping words for chunks. It is applicable only to a word chunking strategy. This value cannot be higher than half the max_chunk_size value.

sentence_overlap number · default: 1

The number of overlapping sentences for chunks. It is applicable only for a sentence chunking strategy. It can be either 1 or 0.

separator_group string

Only applicable to the recursive strategy and required when using it.

Sets a predefined list of separators in the saved chunking settings based on the selected text type. Values can be markdown or plaintext.

Using this parameter is an alternative to manually specifying a custom separators list.

separators string[]

Only applicable to the recursive strategy and required when using it.

A list of strings used as possible split points when chunking text.

Each string can be a plain string or a regular expression (regex) pattern. The system tries each separator in order to split the text, starting from the first item in the list.

After splitting, it attempts to recombine smaller pieces into larger chunks that stay within the max_chunk_size limit, to reduce the total number of chunks generated.

strategy string · default: sentence

The chunking strategy: sentence, word, none or recursive.

  • If strategy is set to recursive, you must also specify:
  • max_chunk_size
  • either separators orseparator_group

Learn more about different chunking strategies in the linked documentation.

Reference
service enum required

The type of service supported for the specified task type. In this case, amazon_sagemaker.

Values: amazon_sagemaker
service_settings object required

Settings used to install the inference model. These settings are specific to the amazon_sagemaker service and service_settings.api you specified.

access_key string required

A valid AWS access key that has permissions to use Amazon SageMaker and access to models for invoking requests.

endpoint_name string required

The name of the SageMaker endpoint.

External docs
api enum required

The API format to use when calling SageMaker. Elasticsearch will convert the POST _inference request to this data format when invoking the SageMaker endpoint.

Values: openai elastic
region string required

The region that your endpoint or Amazon Resource Name (ARN) is deployed in. The list of available regions per model can be found in the Amazon SageMaker documentation.

External docs
secret_key string required

A valid AWS secret key that is paired with the access_key. For information about creating and managing access and secret keys, refer to the AWS documentation.

External docs
similarity enum

Required when api is elastic and task type is text_embedding. The similarity measure used when invoking the text_embedding task type.

Values: cosine dot_product l2_norm
element_type enum

Required when api is elastic and task type is text_embedding. The data type returned by the text embedding model. This value is used when parsing the response back to Elasticsearch data structures.

Values: byte float bit
target_model string

The model ID when calling a multi-model endpoint.

External docs
target_container_hostname string

The container to directly invoke when calling a multi-container endpoint.

External docs
inference_component_name string

The inference component to directly invoke when calling a multi-component endpoint.

External docs
batch_size number · default: 256

The maximum number of inputs in each batch. This value is used by inference ingestion pipelines when processing semantic values. It correlates to the number of times the SageMaker endpoint is invoked (one per batch of input).

dimensions number

The number of dimensions returned by the text embedding models. If this value is not provided, then it is guessed by making invoking the endpoint for the text_embedding task.

task_settings object

Settings to configure the inference task. These settings are specific to the task type and service_settings.api you specified.

custom_attributes string

The AWS custom attributes passed verbatim through to the model running in the SageMaker Endpoint. Values will be returned in the X-elastic-sagemaker-custom-attributes header.

External docs
enable_explanations string

The optional JMESPath expression used to override the EnableExplanations provided during endpoint creation.

External docs
inference_id string

The capture data ID when enabled in the endpoint.

External docs
session_id string

The stateful session identifier for a new or existing session. New sessions will be returned in the X-elastic-sagemaker-new-session-id header. Closed sessions will be returned in the X-elastic-sagemaker-closed-session-id header.

External docs
target_variant string

Specifies the variant when running with multi-variant Endpoints.

External docs

Response

application/json
chunking_settings object

The chunking configuration object. Applies only to the embedding, sparse_embedding and text_embedding task types. Not applicable to the rerank, completion, or chat_completion task types.

max_chunk_size number · default: 250

The maximum size of a chunk in words. This value cannot be lower than 20 (for sentence strategy) or 10 (for word strategy). This value should not exceed the window size for the associated model.

overlap number · default: 100

The number of overlapping words for chunks. It is applicable only to a word chunking strategy. This value cannot be higher than half the max_chunk_size value.

sentence_overlap number · default: 1

The number of overlapping sentences for chunks. It is applicable only for a sentence chunking strategy. It can be either 1 or 0.

separator_group string

Only applicable to the recursive strategy and required when using it.

Sets a predefined list of separators in the saved chunking settings based on the selected text type. Values can be markdown or plaintext.

Using this parameter is an alternative to manually specifying a custom separators list.

separators string[]

Only applicable to the recursive strategy and required when using it.

A list of strings used as possible split points when chunking text.

Each string can be a plain string or a regular expression (regex) pattern. The system tries each separator in order to split the text, starting from the first item in the list.

After splitting, it attempts to recombine smaller pieces into larger chunks that stay within the max_chunk_size limit, to reduce the total number of chunks generated.

strategy string · default: sentence

The chunking strategy: sentence, word, none or recursive.

  • If strategy is set to recursive, you must also specify:
  • max_chunk_size
  • either separators orseparator_group

Learn more about different chunking strategies in the linked documentation.

Reference
service string required

The service type

service_settings object required

Settings specific to the service

task_settings object

Task settings specific to the service and task type

inference_id string required

The inference Id

task_type enum required

The task type

Values: text_embedding completion chat_completion sparse_embedding rerank