Use custom constraints with pipelines

Preview

This product or feature is subject to the "Pre-GA Offerings Terms" in the General Service Terms section of the Service Specific Terms. Pre-GA products and features are available "as is" and might have limited support. For more information, see the launch stage descriptions.

This page shows you how to use Organization Policy Service custom constraints to restrict specific operations on the following Google Cloud resources:

• aiplatform.googleapis.com/PipelineJob

To learn more about Organization Policy, see Custom organization policies.

About organization policies and constraints

The Google Cloud Organization Policy Service gives you centralized, programmatic control over your organization's resources. As the organization policy administrator, you can define an organization policy, which is a set of restrictions called constraints that apply to Google Cloud resources and descendants of those resources in the Google Cloud resource hierarchy. You can enforce organization policies at the organization, folder, or project level.

Organization Policy provides built-in managed constraints for various Google Cloud services. However, if you want more granular, customizable control over the specific fields that are restricted in your organization policies, you can also create custom constraints and use those custom constraints in an organization policy.

Policy inheritance

By default, organization policies are inherited by the descendants of the resources on which you enforce the policy. For example, if you enforce a policy on a folder, Google Cloud enforces the policy on all projects in the folder. To learn more about this behavior and how to change it, refer to Hierarchy evaluation rules.

Benefits

You can use a custom organization policy to allow or deny specific operations on Vertex AI Pipelines resources. For example, if the template URI specified to create a PipelineJob resource fails to satisfy a custom constraint validation set by your organization policy, the request fails, and an error is returned to the caller.

Limitations

Custom organization policies aren't enforced on pipeline runs scheduled using the scheduler API.

Before you begin

1. Set up your project

Sign in to your Google Cloud account. If you're new to Google Cloud, create an account to evaluate how our products perform in real-world scenarios. New customers also get $300 in free credits to run, test, and deploy workloads.

In the Google Cloud console, on the project selector page, select or create a Google Cloud project.

Roles required to select or create a project

• Select a project: Selecting a project doesn't require a specific IAM role—you can select any project that you've been granted a role on.
• Create a project: To create a project, you need the Project Creator role (roles/resourcemanager.projectCreator), which contains the resourcemanager.projects.create permission. Learn how to grant roles.

Note: If you don't plan to keep the resources that you create in this procedure, create a project instead of selecting an existing project. After you finish these steps, you can delete the project, removing all resources associated with the project.

Go to project selector

Verify that billing is enabled for your Google Cloud project.

Enable the Vertex AI, Compute Engine, and Cloud Storage APIs.

Roles required to enable APIs

To enable APIs, you need the Service Usage Admin IAM role (roles/serviceusage.serviceUsageAdmin), which contains the serviceusage.services.enable permission. Learn how to grant roles.

Enable the APIs

Install the Google Cloud CLI.

If you're using an external identity provider (IdP), you must first sign in to the gcloud CLI with your federated identity.

To initialize the gcloud CLI, run the following command:

gcloud init

In the Google Cloud console, on the project selector page, select or create a Google Cloud project.

Roles required to select or create a project

• Select a project: Selecting a project doesn't require a specific IAM role—you can select any project that you've been granted a role on.
• Create a project: To create a project, you need the Project Creator role (roles/resourcemanager.projectCreator), which contains the resourcemanager.projects.create permission. Learn how to grant roles.

Note: If you don't plan to keep the resources that you create in this procedure, create a project instead of selecting an existing project. After you finish these steps, you can delete the project, removing all resources associated with the project.

Go to project selector

Verify that billing is enabled for your Google Cloud project.

Enable the Vertex AI, Compute Engine, and Cloud Storage APIs.

Roles required to enable APIs

To enable APIs, you need the Service Usage Admin IAM role (roles/serviceusage.serviceUsageAdmin), which contains the serviceusage.services.enable permission. Learn how to grant roles.

Enable the APIs

Install the Google Cloud CLI.

If you're using an external identity provider (IdP), you must first sign in to the gcloud CLI with your federated identity.

To initialize the gcloud CLI, run the following command:

gcloud init

1. Get your organization resource ID.
2. Define and compile a pipeline that you can use to test the custom constraint.

Required roles

To get the permissions that you need to manage organization policies, ask your administrator to grant you the following IAM roles:

• To manage organization policies: Organization Policy Administrator (roles/orgpolicy.policyAdmin) on the organization resource
• Create or update an ML pipeline: Vertex AI Admin (roles/aiplatform.admin) or Vertex AI User (roles/aiplatform.user) on the project resource

For more information about granting roles, see Manage access to projects, folders, and organizations.

You might also be able to get the required permissions through custom roles or other predefined roles.

Set up a custom constraint

A custom constraint is defined in a YAML file by the resources, methods, conditions, and actions that are supported by the service on which you are enforcing the organization policy. Conditions for your custom constraints are defined using Common Expression Language (CEL). For more information about how to build conditions in custom constraints using CEL, see the CEL section of Creating and managing custom constraints.

Console

To create a custom constraint, do the following:

1. In the Google Cloud console, go to the Organization policies page.

   Go to Organization policies

2. From the project picker, select the project that you want to set the organization policy for.
3. Click Custom constraint.
4. In the Display name box, enter a human-readable name for the constraint. This name is used in error messages and can be used for identification and debugging. Don't use PII or sensitive data in display names because this name could be exposed in error messages. This field can contain up to 200 characters.
5. In the Constraint ID box, enter the name that you want for your new custom constraint. A custom constraint can only contain letters (including upper and lowercase) or numbers, for example custom.disableGkeAutoUpgrade. This field can contain up to 70 characters, not counting the prefix (custom.), for example, organizations/123456789/customConstraints/custom. Don't include PII or sensitive data in your constraint ID, because it could be exposed in error messages.
6. In the Description box, enter a human-readable description of the constraint. This description is used as an error message when the policy is violated. Include details about why the policy violation occurred and how to resolve the policy violation. Don't include PII or sensitive data in your description, because it could be exposed in error messages. This field can contain up to 2000 characters.
7. In the Resource type box, select the name of the Google Cloud REST resource containing the object and field that you want to restrict—for example, container.googleapis.com/NodePool. Most resource types support up to 20 custom constraints. If you attempt to create more custom constraints, the operation fails.
8. Under Enforcement method, select whether to enforce the constraint on a REST CREATE method or on both CREATE and UPDATE methods. If you enforce the constraint with the UPDATE method on a resource that violates the constraint, changes to that resource are blocked by the organization policy unless the change resolves the violation.

Not all Google Cloud services support both methods. To see supported methods for each service, find the service in Supported services.

9. To define a condition, click Edit condition.
1. In the Add condition panel, create a CEL condition that refers to a supported service resource, for example, resource.management.autoUpgrade == false. This field can contain up to 1000 characters. For details about CEL usage, see Common Expression Language. For more information about the service resources you can use in your custom constraints, see Custom constraint supported services.
2. Click Save.
10. Under Action, select whether to allow or deny the evaluated method if the condition is met.

The deny action means that the operation to create or update the resource is blocked if the condition evaluates to true.

The allow action means that the operation to create or update the resource is permitted only if the condition evaluates to true. Every other case except ones explicitly listed in the condition is blocked.

11. Click Create constraint.

When you have entered a value into each field, the equivalent YAML configuration for this custom constraint appears on the right.

gcloud

1. To create a custom constraint, create a YAML file using the following format:

      name: organizations/ORGANIZATION_ID/customConstraints/CONSTRAINT_NAME
      resourceTypes:
      - RESOURCE_NAME
      methodTypes:
      - CREATE
      condition: "CONDITION"
      actionType: ACTION
      displayName: DISPLAY_NAME
      description: DESCRIPTION
      

Replace the following:

• ORGANIZATION_ID: your organization ID, such as 123456789.
• CONSTRAINT_NAME: the name that you want for your new custom constraint. A custom constraint can only contain letters (including upper and lowercase) or numbers, for example, custom.denyPipelineTemplate. This field can contain up to 70 characters.
• RESOURCE_NAME: the fully qualified name of the Google Cloud resource containing the object and field that you want to restrict. For example, aiplatform.googleapis.com/PipelineJob/resource.templateUri.
• CONDITION: a CEL condition that is written against a representation of a supported service resource. This field can contain up to 1000 characters. For example, "resource.templateUri.contains("test")".

For more information about the resources available to write conditions against, see Supported resources.

• ACTION: the action to take if the condition is met. Possible values are ALLOW and DENY.

The allow action means that if the condition evaluates to true, the operation to create or update the resource is permitted. This also means that every other case except the one explicitly listed in the condition is blocked.

The deny action means that if the condition evaluates to true, the operation to create or update the resource is blocked.

• DISPLAY_NAME: a human-friendly name for the constraint. This field can contain up to 200 characters.
• DESCRIPTION: a human-friendly description of the constraint to display as an error message when the policy is violated. This field can contain up to 2000 characters.
2. After you have created the YAML file for a new custom constraint, you must set it up to make it available for organization policies in your organization. To set up a custom constraint, use the gcloud org-policies set-custom-constraint command:

        gcloud org-policies set-custom-constraint CONSTRAINT_PATH
      

Replace CONSTRAINT_PATH with the full path to your custom constraint file. For example, /home/user/customconstraint.yaml.

After this operation is complete, your custom constraints are available as organization policies in your list of Google Cloud organization policies.

3. To verify that the custom constraint exists, use the gcloud org-policies list-custom-constraints command:

      gcloud org-policies list-custom-constraints --organization=ORGANIZATION_ID
      

Replace ORGANIZATION_ID with the ID of your organization resource.

For more information, see Viewing organization policies.

Enforce a custom organization policy

You can enforce a constraint by creating an organization policy that references it, and then applying that organization policy to a Google Cloud resource.

Console

1. In the Google Cloud console, go to the Organization policies page.

   Go to Organization policies

2. From the project picker, select the project that you want to set the organization policy for.
3. From the list on the Organization policies page, select your constraint to view the Policy details page for that constraint.
4. To configure the organization policy for this resource, click Manage policy.
5. On the Edit policy page, select Override parent's policy.
6. Click Add a rule.
7. In the Enforcement section, select whether this organization policy is enforced or not.
8. Optional: To make the organization policy conditional on a tag, click Add condition. Note that if you add a conditional rule to an organization policy, you must add at least one unconditional rule or the policy cannot be saved. For more information, see Setting an organization policy with tags.
9. Click Test changes to simulate the effect of the organization policy. For more information, see Test organization policy changes with Policy Simulator.
10. To enforce the organization policy in dry-run mode, click Set dry run policy. For more information, see Create an organization policy in dry-run mode.
11. After you verify that the organization policy in dry-run mode works as intended, set the live policy by clicking Set policy.

gcloud

1. To create an organization policy with boolean rules, create a policy YAML file that references the constraint:

        name: projects/PROJECT_ID/policies/CONSTRAINT_NAME
        spec:
          rules:
          - enforce: true
        
        dryRunSpec:
          rules:
          - enforce: true
        
      

Replace the following:

• PROJECT_ID: the project that you want to enforce your constraint on.
• CONSTRAINT_NAME: the name you defined for your custom constraint. For example, custom.denyPipelineTemplate.
2. To enforce the organization policy in dry-run mode, run the following command with the dryRunSpec flag:

        gcloud org-policies set-policy POLICY_PATH \
          --update-mask=dryRunSpec
      

Replace POLICY_PATH with the full path to your organization policy YAML file. The policy requires up to 15 minutes to take effect.

3. After you verify that the organization policy in dry-run mode works as intended, set the live policy with the org-policies set-policy command and the spec flag:

        gcloud org-policies set-policy POLICY_PATH \
          --update-mask=spec
      

Replace POLICY_PATH with the full path to your organization policy YAML file. The policy requires up to 15 minutes to take effect.

Test the custom organization policy

The following example creates a custom constraint and policy that don't allow pipeline runs to be created by specifying a template URI that contains "test".

Before you begin, you must know the following:

• Your organization ID
• A project ID

Create the constraint

1. Save the following file as constraint-validate-pipeline-template-uri.yaml:

   name: organizations/ORGANIZATION_ID/customConstraints/custom.denyPipelineTemplate
   resourceTypes:
   - resource.templateUri
   methodTypes:
     - CREATE
     condition: "resource.templateUri.contains("test")"
     actionType: DENY
     displayName: Deny pipeline runs if the template URI contains 'test'
     description: Deny the creation of a new pipeline run if it's based on a template URI containing 'test'
   

   This defines a constraint where the pipeline template URI can't contain test.

   1. Apply the constraint:

      gcloud org-policies set-custom-constraint ~/constraint-validate-pipeline-template-uri.yaml
      

   2. Verify that the constraint exists:

      gcloud org-policies list-custom-constraints --organization=ORGANIZATION_ID
      

      The output is similar to the following:

      CUSTOM_CONSTRAINT                          ACTION_TYPE  METHOD_TYPES   RESOURCE_TYPES          DISPLAY_NAME
      custom.denyPipelineTemplate                DENY         CREATE         resource.templateUri    Deny pipeline runs if the template URI contains 'test'
      ...
      

Create the policy

1. Save the following file as policy-validate-pipeline-template-uri.yaml:

     name: projects/PROJECT_ID/policies/custom.denyPipelineTemplate
     spec:
       rules:
       - enforce: true
   

   Replace PROJECT_ID with your project ID.

2. Apply the policy:

     gcloud org-policies set-policy ~/policy-validate-pipeline-template-uri.yaml
   

3. Verify that the policy exists:

     gcloud org-policies list --project=PROJECT_ID
   

   The output is similar to the following:

     CONSTRAINT                    LIST_POLICY    BOOLEAN_POLICY    ETAG
     custom.denyPipelineTemplate   -              SET               COCsm5QGENiXi2E=
   

   After you apply the policy, wait for about two minutes for Google Cloud to start enforcing the policy.

Test the policy

Try to create an ML pipeline with the template URI containing test.

REST

To create a PipelineJob resource, send a POST request by using the pipelineJobs/create method.

Before using any of the request data, make the following replacements:

• LOCATION: The region where you want to create the pipeline run. For more information about the regions where Vertex AI Pipelines is available, see the Vertex AI Pipelines locations guide.
• PROJECT_ID: The Google Cloud project where you want to create the pipeline run.
• DISPLAY_NAME: The name of the pipeline run. This will be displayed in the Google Cloud console.

HTTP method and URL:

POST https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/pipelineJobs

Request JSON body:

{
  "displayName":"DISPLAY_NAME",
  "templateUri":"test_pipeline_template.json"
}

To send your request, choose one of these options:

curl

Note: The following command assumes that you have logged in to the gcloud CLI with your user account by running gcloud init or gcloud auth login , or by using Cloud Shell, which automatically logs you into the gcloud CLI . You can check the currently active account by running gcloud auth list.

Save the request body in a file named request.json, and execute the following command:

curl -X POST \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     -H "Content-Type: application/json; charset=utf-8" \
     -d @request.json \
     "https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/pipelineJobs"

PowerShell

Note: The following command assumes that you have logged in to the gcloud CLI with your user account by running gcloud init or gcloud auth login . You can check the currently active account by running gcloud auth list.

Save the request body in a file named request.json, and execute the following command:

$cred = gcloud auth print-access-token
$headers = @{ "Authorization" = "Bearer $cred" }

Invoke-WebRequest `
    -Method POST `
    -Headers $headers `
    -ContentType: "application/json; charset=utf-8" `
    -InFile request.json `
    -Uri "https://LOCATION-aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/LOCATION/pipelineJobs" | Select-Object -Expand Content

You should receive a JSON response similar to the following:

{
  "error": {
    "code": 400,
    "message": "Operation denied by org policy on resource 'projects/PROJECT_ID/locations/LOCATION': [\"customConstraints/custom.denyPipelineTemplate\": \"Deny the creation of a new pipeline run if it's based on a template URI containing 'test'\"]",
    "status": "FAILED_PRECONDITION",
    "details": [
      {
        "@type": "type.googleapis.com/google.rpc.ErrorInfo",
        "reason": "CUSTOM_ORG_POLICY_VIOLATION",
        "domain": "googleapis.com",
        "metadata": {
          "service": "aiplatform.googleapis.com",
          "customConstraints": "customConstraints/custom.denyPipelineTemplate"
        }
      }
    ]
  }
}

Vertex AI Pipelines supported resources

The following table lists the Vertex AI Pipelines resources that you can reference in custom constraints.

Resource	Field
aiplatform.googleapis.com/PipelineJob	resource.displayName
	resource.encryptionSpec.kmsKeyName
	resource.network
	resource.pipelineSpec
	resource.preflightValidations
	resource.pscInterfaceConfig.networkAttachment
	resource.reservedIpRanges
	resource.runtimeConfig.failurePolicy
	resource.runtimeConfig.gcsOutputDirectory
	resource.runtimeConfig.inputArtifacts[*].artifactId
	resource.runtimeConfig.parameterValues[*].boolValue
	resource.runtimeConfig.parameterValues[*].listValue.values
	resource.runtimeConfig.parameterValues[*].nullValue
	resource.runtimeConfig.parameterValues[*].numberValue
	resource.runtimeConfig.parameterValues[*].stringValue
	resource.runtimeConfig.parameterValues[*].structValue
	resource.serviceAccount
	resource.templateUri

What's next

• Learn more about Organization Policy Service.
• Learn more about how to create and manage organization policies.
• See the full list of managed organization policy constraints.