guardrails

Creates, updates, deletes or gets a guardrail resource or lists guardrails in a region

Overview

Name	guardrails
Type	Resource
Description	Definition of AWS::Bedrock::Guardrail Resource Type
Id	awscc.bedrock.guardrails

Fields

get (all properties)

Name	Datatype	Description
blocked_input_messaging	string	Messaging for when violations are detected in text
blocked_outputs_messaging	string	Messaging for when violations are detected in text
▶content_policy_config	object	Content policy config for a guardrail.
▶contextual_grounding_policy_config	object	Contextual grounding policy config for a guardrail.
created_at	string	Time Stamp
▶cross_region_config	object	The system-defined guardrail profile that you’re using with your guardrail
description	string	Description of the guardrail or its version
failure_recommendations	array	List of failure recommendations
guardrail_arn	string	Arn representation for the guardrail
guardrail_id	string	Unique id for the guardrail
kms_key_arn	string	The KMS key with which the guardrail was encrypted at rest
name	string	Name of the guardrail
▶sensitive_information_policy_config	object	Sensitive information policy config for a guardrail.
status	string	Status of the guardrail
status_reasons	array	List of status reasons
▶tags	array	List of Tags
▶topic_policy_config	object	Topic policy config for a guardrail.
updated_at	string	Time Stamp
version	string	Guardrail version
▶word_policy_config	object	Word policy config for a guardrail.
region	string	AWS region.

list (identifiers only)

Name	Datatype	Description
guardrail_arn	string	Arn representation for the guardrail
region	string	AWS region.

For more information, see AWS::Bedrock::Guardrail.

Methods

Name	Resource	Accessible by	Required Params
create_resource	guardrails	INSERT	Name, BlockedInputMessaging, BlockedOutputsMessaging, region
delete_resource	guardrails	DELETE	Identifier, region
update_resource	guardrails	UPDATE	Identifier, PatchDocument, region
list_resources	guardrails_list_only	SELECT	region
get_resource	guardrails	SELECT	Identifier, region

SELECT examples

get (all properties)

Gets all properties from an individual guardrail.

SELECT
  region,
  blocked_input_messaging,
  blocked_outputs_messaging,
  content_policy_config,
  contextual_grounding_policy_config,
  created_at,
  cross_region_config,
  description,
  failure_recommendations,
  guardrail_arn,
  guardrail_id,
  kms_key_arn,
  name,
  sensitive_information_policy_config,
  status,
  status_reasons,
  tags,
  topic_policy_config,
  updated_at,
  version,
  word_policy_config
FROM awscc.bedrock.guardrails
WHERE
  region = 'us-east-1' AND
  Identifier = '{{ guardrail_arn }}';

list (identifiers only)

Lists all guardrails in a region.

SELECT
  region,
  guardrail_arn
FROM awscc.bedrock.guardrails_list_only
WHERE
  region = 'us-east-1';

INSERT example

Use the following StackQL query and manifest file to create a new guardrail resource, using stack-deploy.

Required Properties

/*+ create */
INSERT INTO awscc.bedrock.guardrails (
  BlockedInputMessaging,
  BlockedOutputsMessaging,
  Name,
  region
)
SELECT
  '{{ blocked_input_messaging }}',
  '{{ blocked_outputs_messaging }}',
  '{{ name }}',
  '{{ region }}';

All Properties

/*+ create */
INSERT INTO awscc.bedrock.guardrails (
  BlockedInputMessaging,
  BlockedOutputsMessaging,
  ContentPolicyConfig,
  ContextualGroundingPolicyConfig,
  CrossRegionConfig,
  Description,
  KmsKeyArn,
  Name,
  SensitiveInformationPolicyConfig,
  Tags,
  TopicPolicyConfig,
  WordPolicyConfig,
  region
)
SELECT
  '{{ blocked_input_messaging }}',
  '{{ blocked_outputs_messaging }}',
  '{{ content_policy_config }}',
  '{{ contextual_grounding_policy_config }}',
  '{{ cross_region_config }}',
  '{{ description }}',
  '{{ kms_key_arn }}',
  '{{ name }}',
  '{{ sensitive_information_policy_config }}',
  '{{ tags }}',
  '{{ topic_policy_config }}',
  '{{ word_policy_config }}',
  '{{ region }}';

Manifest

version: 1
name: stack name
description: stack description
providers:
  - aws
globals:
  - name: region
    value: '{{ vars.AWS_REGION }}'
resources:
  - name: guardrail
    props:
      - name: blocked_input_messaging
        value: '{{ blocked_input_messaging }}'
      - name: blocked_outputs_messaging
        value: '{{ blocked_outputs_messaging }}'
      - name: content_policy_config
        value:
          filters_config:
            - type: '{{ type }}'
              input_strength: '{{ input_strength }}'
              output_strength: null
              input_modalities:
                - '{{ input_modalities[0] }}'
              output_modalities:
                - null
              input_action: '{{ input_action }}'
              output_action: null
              input_enabled: '{{ input_enabled }}'
              output_enabled: '{{ output_enabled }}'
          content_filters_tier_config:
            tier_name: '{{ tier_name }}'
      - name: contextual_grounding_policy_config
        value:
          filters_config:
            - type: '{{ type }}'
              threshold: null
              action: '{{ action }}'
              enabled: '{{ enabled }}'
      - name: cross_region_config
        value:
          guardrail_profile_arn: '{{ guardrail_profile_arn }}'
      - name: description
        value: '{{ description }}'
      - name: kms_key_arn
        value: '{{ kms_key_arn }}'
      - name: name
        value: '{{ name }}'
      - name: sensitive_information_policy_config
        value:
          pii_entities_config:
            - type: '{{ type }}'
              action: '{{ action }}'
              input_action: null
              output_action: null
              input_enabled: '{{ input_enabled }}'
              output_enabled: '{{ output_enabled }}'
          regexes_config:
            - name: '{{ name }}'
              description: '{{ description }}'
              pattern: '{{ pattern }}'
              action: null
              input_action: null
              output_action: null
              input_enabled: '{{ input_enabled }}'
              output_enabled: '{{ output_enabled }}'
      - name: tags
        value:
          - key: '{{ key }}'
            value: '{{ value }}'
      - name: topic_policy_config
        value:
          topics_config:
            - name: '{{ name }}'
              definition: '{{ definition }}'
              examples:
                - '{{ examples[0] }}'
              type: '{{ type }}'
              input_action: '{{ input_action }}'
              output_action: null
              input_enabled: '{{ input_enabled }}'
              output_enabled: '{{ output_enabled }}'
          topics_tier_config:
            tier_name: '{{ tier_name }}'
      - name: word_policy_config
        value:
          words_config:
            - text: '{{ text }}'
              input_action: '{{ input_action }}'
              output_action: null
              input_enabled: '{{ input_enabled }}'
              output_enabled: '{{ output_enabled }}'
          managed_word_lists_config:
            - type: '{{ type }}'
              input_action: null
              output_action: null
              input_enabled: '{{ input_enabled }}'
              output_enabled: '{{ output_enabled }}'

UPDATE example

Use the following StackQL query and manifest file to update a guardrail resource, using stack-deploy.

/*+ update */
UPDATE awscc.bedrock.guardrails
SET PatchDocument = string('{{ {
    "BlockedInputMessaging": blocked_input_messaging,
    "BlockedOutputsMessaging": blocked_outputs_messaging,
    "ContentPolicyConfig": content_policy_config,
    "ContextualGroundingPolicyConfig": contextual_grounding_policy_config,
    "CrossRegionConfig": cross_region_config,
    "Description": description,
    "KmsKeyArn": kms_key_arn,
    "Name": name,
    "SensitiveInformationPolicyConfig": sensitive_information_policy_config,
    "Tags": tags,
    "TopicPolicyConfig": topic_policy_config,
    "WordPolicyConfig": word_policy_config
} | generate_patch_document }}')
WHERE
  region = '{{ region }}' AND
  Identifier = '{{ guardrail_arn }}';

DELETE example

/*+ delete */
DELETE FROM awscc.bedrock.guardrails
WHERE
  Identifier = '{{ guardrail_arn }}' AND
  region = 'us-east-1';

Permissions

To operate on the guardrails resource, the following permissions are required:

Create

bedrock:CreateGuardrail,
bedrock:GetGuardrail,
kms:DescribeKey,
kms:CreateGrant,
kms:GenerateDataKey,
kms:Decrypt,
bedrock:TagResource,
bedrock:ListTagsForResource

Read

bedrock:GetGuardrail,
kms:Decrypt,
bedrock:ListTagsForResource

Update

bedrock:UpdateGuardrail,
bedrock:GetGuardrail,
bedrock:ListTagsForResource,
bedrock:TagResource,
bedrock:UntagResource,
kms:DescribeKey,
kms:CreateGrant,
kms:GenerateDataKey,
kms:Decrypt

Delete

bedrock:DeleteGuardrail,
bedrock:GetGuardrail,
kms:Decrypt,
kms:RetireGrant

List

bedrock:ListGuardrails