View a markdown version of this page

Product columns - AWS Data Exports
Service-specific product attributes

Product columns

Product columns contain data about the product that is being charged in the line item.

Column name Description Data type
product

A map column containing key-value pairs of multiple product attributes and their values for a given line item.

A product attribute only appears in the map column if it has a value that applies to the specific line item.

Note

Any product column that appeared in legacy CUR, but is not part of the CUR 2.0 static schema, appears in this map column.

The keys of this column can be queried as individual columns by using the dot operator. For more information, see Data query.

For some AWS services, the map includes additional service-specific attributes. For the list of these attributes and the services that populate them, see Service-specific product attributes.

 map <string, string>
product_comment

A comment regarding the product.

 string
product_fee_code

The code that refers to the fee.

 string
product_fee_description

The description for the product fee.

 string
product_from_location

Describes the location where the usage originated from.

 string
product_from_location_type

Describes the location type where the usage originated from.

 string
product_from_region_code

Describes the source Region code for the AWS service.

 string
product_instanceSKU

The SKU of the product instance

 string
product_instance_family

Describes your Amazon EC2 instance family. Amazon EC2 provides you with a large number of options across 10 different instance types, each with one or more size options, organized into distinct instance families optimized for different types of applications.

 string
product_instance_type

Describes the instance type, size, and family, which define the CPU, networking, and storage capacity of your instance.

 string
product_location

Describes the Region that your Amazon S3 bucket resides in.

 string
product_location_type

Describes the endpoint of your task.

 string
product_operation

Describes the specific AWS operation that this line item covers.

 string
product_pricing_unit

The smallest billing unit for an AWS service. For example, 0.01c per API call.

 string
product_product_family

The category for the type of product. For Amazon Bedrock line items, this is populated with the unified value Amazon Bedrock, so that all Amazon Bedrock usage can be identified under a single product family.

 string
product_region_code

A Region is a physical location around the world where data centers are clustered. AWS calls each group of logical data centers an Availability Zone (AZ). Each AWS Region consists of multiple, isolated, and physically separate AZs within a geographical area. The Region code attribute has the same name as an AWS Region, and specifies where the AWS service is available.

 string
product_sku

A unique code for a product. The SKU is created by combining the ProductCode, UsageType, and Operation. For size-flexible RIs, the SKU uses the instance that was used. For example, if you used a t2.micro instance and AWS applied a t2.small RI discount to the usage, the line item SKU is created with the t2.micro.

 string
product_servicecode

This identifies the specific AWS service to the customer as a unique short abbreviation.

 string
product_to _location_type

Describes the destination location of the service usage.

 string
product_to_location

Describes the location usage destination.

 string
product_to_region_code

Describes the source Region code for the AWS service.

 string
product_usagetype

Describes the usage details of the line item.

 string

Service-specific product attributes

For certain AWS services, the product map column includes additional standardized attributes that describe service-specific cost drivers beyond the static schema columns. These attributes are stored as keys within the product map rather than as static top-level columns, and you can query each one by using the dot operator (for example, product.provider). For more information about querying map keys, see Data query.

These attributes let you filter, group, and compare costs along service-specific dimensions without parsing free-text description fields. For example, for Amazon Bedrock usage you can group costs by product.provider to compare spending across model providers, or by product.inference_type to separate input token costs from output token costs.

The Applies to column indicates which AWS service populates each attribute.

Map key Description Example values Applies to
provider

The provider of the foundation model.

Amazon, Anthropic, Meta, Cohere

Amazon Bedrock
model

The model name, corresponding to the model name shown in the Amazon Bedrock console Model catalog.

Claude 3.5 Sonnet, Nova Pro, Llama 3

Amazon Bedrock
inference_type

The type of usage being metered (the token type).

input tokens, output tokens, image generation

Amazon Bedrock
feature

The inference serving mode, describing how the request is processed.

On-Demand, Batch

Amazon Bedrock