Amazon S3 - Amazon Kendra
Supported featuresPrerequisitesConnection instructions

Amazon S3

Amazon S3 is an object storage service that stores data as objects within buckets. You can use Amazon Kendra to index your Amazon S3 bucket repository of documents.

Warning

Amazon Kendra doesn't use a bucket policy that grants permissions to an Amazon Kendra principal to interact with an S3 bucket. Instead, it uses IAM roles. Make sure that Amazon Kendra isn't included as a trusted member in your bucket policy to avoid any data security issues in accidentally granting permissions to arbitrary principals. However, you can add a bucket policy to use an Amazon S3 bucket across different accounts. For more information, see Policies to use Amazon S3 across accounts (within the S3 IAM roles tab, under IAM roles for data sources). For information about IAM roles for S3 data sources, see IAM roles.

Note

Amazon Kendra now supports an upgraded Amazon S3 connector.

The console has been automatically upgraded for you. Any new connectors you create in the console will use the upgraded architecture. If you use the API, you must now use the TemplateConfiguration object instead of the S3DataSourceConfiguration object to configure your connector.

Connectors configured using the older console and API architecture will continue to function as configured. However, you won’t be able to edit or update them. If you want to edit or update your connector configuration, you must create a new connector.

We recommended migrating your connector workflow to the upgraded version. Support for connectors configured using the older architecture is scheduled to end by June 2024.

You can connect to your Amazon S3 data source using the the Amazon Kendra console or the TemplateConfiguration API.

Note

To generate a sync status report for your Amazon S3 data source, see Troubleshooting data sources.

For troubleshooting your Amazon Kendra S3 data source connector, see Troubleshooting data sources.

Topics

Supported features

  • Field mappings

  • User access control

  • Inclusion/exclusion filters

  • Full and incremental content syncs

  • Virtual private cloud (VPC)

Prerequisites

Before you can use Amazon Kendra to index your S3 data source, make these changes in your S3 and AWS accounts.

In S3, make sure you have:

  • Copied the name of your Amazon S3 bucket.

    Note

    Your bucket must be in the same region as your Amazon Kendra index and your index must have permission to access the bucket that contains your documents.

  • Checked each document is unique in S3 and across other data sources you plan to use for the same index. Each data source that you want to use for an index must not contain the same document across the data sources. Document IDs are global to an index and must be unique per index.

In your AWS account, make sure you have:

If you don’t have an existing IAM role, you can use the console to create a new IAM role when you connect your S3 data source to Amazon Kendra. If you are using the API, you must provide the ARN of an existing IAM role and an index ID.

Connection instructions

To connect Amazon Kendra to your S3 data source, you must provide the necessary details of your S3 data source so that Amazon Kendra can access your data. If you have not yet configured S3 for Amazon Kendra, see Prerequisites.

Console

To connect Amazon Kendra to Amazon S3

  1. Sign in to the AWS Management Console and open the Amazon Kendra console.

  2. From the left navigation pane, choose Indexes and then choose the index you want to use from the list of indexes.

    Note

    You can choose to configure or edit your User access control settings under Index settings.

  3. On the Getting started page, choose Add data source.

  4. On the Add data source page, choose S3 connector, and then choose Add connector. If using version 2 (if applicable), choose S3 connector with the "V2.0" tag.

  5. On the Specify data source details page, enter the following information:

    1. In Name and description, for Data source name—Enter a name for your data source. You can include hyphens but not spaces.

    2. (Optional) Description—Enter an optional description for your data source.

    3. In Default language—Choose a language to filter your documents for the index. Unless you specify otherwise, the language defaults to English. Language specified in the document metadata overrides the selected language.

    4. In Tags, for Add new tag—Include optional tags to search and filter your resources or track your AWS costs.

    5. Choose Next.

  6. On the Define access and security page, enter the following optional information:

    1. IAM role—Choose an existing IAM role or create a new IAM role to access your repository credentials and index content.

      Note

      IAM roles used for indexes cannot be used for data sources. If you are unsure if an existing role is used for an index or FAQ, choose Create a new role to avoid errors.

    2. Virtual Private Cloud (VPC)—You can choose to use a VPC. If so, you must add Subnets and VPC security groups.

    3. Choose Next.

  7. On the Configure sync settings page, enter the following information:

    1. For Data source location—Specify the path to the Amazon S3 bucket where your data is stored. Select Browse S3 to choose your S3 bucket.

    2. For Maximum file size—Specify a limit in MB to only crawl files under this limit. The maximum file size Amazon Kendra can allow is 50 MB.

    3. For (Optional) Metadata files prefix folder location—Specify the path to the folder in which your fields/attributes and other document metadata is stored. Select Browse S3 to locate your metadata folder.

    4. For (Optional) Access control list configuration file location—Specify the path to the file that contains a JSON structure of your users and their access to documents. Select Browse S3 to locate your ACL file.

    5. (Optional) Select decryption key—Select to use a decryption key. You can choose to use an existing AWS KMS key.

    6. For (Optional) Additional configuration—Add patterns to include or exclude certain files. All paths are relative to the data source location S3 bucket.

    7. Sync mode—Choose how you want to update your index when your data source content changes. When you sync your data source with Amazon Kendra for the first time, all content is crawled and indexed by default. You must run a full sync of your data if your initial sync failed, even if you don't choose full sync as your sync mode option.

      • Full sync: Freshly index all content, replacing existing content each time your data source syncs with your index.

      • New, modified, deleted sync: Index only new, modified, and deleted content each time your data source syncs with your index. Amazon Kendra can use your data source's mechanism for tracking content changes and index content that changed since the last sync.

    8. In Sync run schedule, for Frequency—Choose how often to sync your data source content and update your index.

    9. Choose Next.

  8. On the Set field mappings page, enter the following optional information:

    1. Default field mappings—Select from the Amazon Kendra generated default data source fields you want to map to your index.

    2. Add field—Choose to add custom data source fields to create an index field name to map to and the field data type.

    3. Choose Next.

  9. On the Review and create page, check that the information you have entered is correct and then select Add data source. You can also choose to edit your information from this page. Your data source will appear on the Data sources page after the data source has been added successfully.

API

To connect Amazon Kendra to Amazon S3

You must specify a JSON of the data source schema using the TemplateConfiguration API. You must provide the following information:

  • Data source—Specify the data source type as S3 when you use the TemplateConfiguration JSON schema. Also specify the data source as TEMPLATE when you call the CreateDataSource API.

  • BucketName—The name of the bucket that contains the documents.

  • Sync mode—Specify how Amazon Kendra should update your index when your data source content changes. When you sync your data source with Amazon Kendra for the first time, all content is crawled and indexed by default. You must run a full sync of your data if your initial sync failed, even if you don't choose full sync as your sync mode option. You can choose between:

    • FORCED_FULL_CRAWL to freshly index all content, replacing existing content each time your data source syncs with your index.

    • FULL_CRAWL to index only new, modified, and deleted content each time your data source syncs with your index. Amazon Kendra can use your data source’s mechanism for tracking content changes and index content that changed since the last sync.

  • IAM role—Specify RoleArn when you call CreateDataSource to provide an IAM role with permissions to access your Secrets Manager secret and to call the required public APIs for the S3 connector and Amazon Kendra. For more information, see IAM roles for S3 data sources.

You can also add the following optional features:

  • Virtual Private Cloud (VPC)—Specify VpcConfiguration when you call CreateDataSource. For more information, see Configuring Amazon Kendra to use an Amazon VPC.

  • Inclusion and exclusion filters—Specify whether to include or exclude certain file names, file types, file paths. You use glob patterns (patterns that can expand a wildcard pattern into a list of path names that match the given pattern). For examples, see Use of Exclude and Include Filters in the AWS CLI Command Reference.

  • Document metadata and access control configuration—Add document metadata and access control files that contain information such as the source URI, document author, or custom document attributes/fields, and your users and which documents they can access. Each metadata file contains metadata about a single document.

  • Field mappings—Choose to map your S3 data source fields to your Amazon Kendra index fields. For more information, see Mapping data source fields.

    Note

    The document body field or the document body equivalent for your documents is required in order for Amazon Kendra to search your documents. You must map your document body field name in your data source to the index field name _document_body. All other fields are optional.

For a list of other important JSON keys to configure, see S3 template schema.

Learn more

To learn more about integrating Amazon Kendra with your S3 data source, see: