Controle de acesso da Vertex AI com o IAM

Nesta página, descrevemos como usar o Identity and Access Management (IAM) para gerenciar o acesso aos recursos da Vertex AI. Para gerenciar o acesso aos recursos do Vertex AI Workbench, consulte as páginas de controle de acesso para notebooks gerenciados ou notebooks gerenciados pelo usuário.

A Vertex AI usa o IAM para gerenciar o acesso aos recursos. É possível gerenciar o acesso no nível do projeto ou do recurso. Para conceder acesso a recursos no nível do projeto, atribua um ou mais papéis a um principal (usuário, grupo ou conta de serviço). Para conceder acesso a um recurso específico, defina uma política do IAM nesse recurso: o recurso precisa ser compatível com políticas no nível do recurso. A política define quais papéis são atribuídos a quais principais.

Há diferentes tipos de papéis do IAM que podem ser usados na Vertex AI:

• Os papéis predefinidos permitem conceder um conjunto de permissões relacionadas aos recursos da Vertex AI no nível do projeto.

• Os papéis básicos (proprietário, editor e visualizador) fornecem controle de acesso aos recursos da Vertex AI no nível do projeto e são comuns a todos os serviços do Google Cloud.

• Os papéis personalizados permitem escolher um conjunto específico de permissões, criar seu próprio papel com elas e concedê-lo aos usuários da organização.

Para adicionar, atualizar ou remover esses papéis do projeto da Vertex AI, consulte a documentação sobre como conceder, alterar e revogar acesso.

Papéis predefinidos da Vertex AI

Papel	Permissões
Administrador da Vertex AI (roles/aiplatform.admin) Concede acesso total a todos os recursos da Vertex AI Contém 3 permissões de proprietário	aiplatform.* aiplatform.annotationSpecs.create aiplatform.annotationSpecs.delete aiplatform.annotationSpecs.get aiplatform.annotationSpecs.list aiplatform.annotationSpecs.update aiplatform.annotations.create aiplatform.annotations.delete aiplatform.annotations.get aiplatform.annotations.list aiplatform.annotations.update aiplatform.artifacts.create aiplatform.artifacts.delete aiplatform.artifacts.get aiplatform.artifacts.list aiplatform.artifacts.update aiplatform.batchPredictionJobs.cancel aiplatform.batchPredictionJobs.create aiplatform.batchPredictionJobs.delete aiplatform.batchPredictionJobs.get aiplatform.batchPredictionJobs.list aiplatform.contexts.addContextArtifactsAndExecutions aiplatform.contexts.addContextChildren aiplatform.contexts.create aiplatform.contexts.delete aiplatform.contexts.get aiplatform.contexts.list aiplatform.contexts.queryContextLineageSubgraph aiplatform.contexts.update aiplatform.customJobs.cancel aiplatform.customJobs.create aiplatform.customJobs.delete aiplatform.customJobs.get aiplatform.customJobs.list aiplatform.dataItems.create aiplatform.dataItems.delete aiplatform.dataItems.get aiplatform.dataItems.list aiplatform.dataItems.update aiplatform.dataLabelingJobs.cancel aiplatform.dataLabelingJobs.create aiplatform.dataLabelingJobs.delete aiplatform.dataLabelingJobs.get aiplatform.dataLabelingJobs.list aiplatform.datasets.create aiplatform.datasets.delete aiplatform.datasets.export aiplatform.datasets.get aiplatform.datasets.import aiplatform.datasets.list aiplatform.datasets.update aiplatform.deploymentResourcePools.create aiplatform.deploymentResourcePools.delete aiplatform.deploymentResourcePools.get aiplatform.deploymentResourcePools.list aiplatform.deploymentResourcePools.queryDeployedModels aiplatform.deploymentResourcePools.update aiplatform.edgeDeploymentJobs.create aiplatform.edgeDeploymentJobs.delete aiplatform.edgeDeploymentJobs.get aiplatform.edgeDeploymentJobs.list aiplatform.edgeDeviceDebugInfo.get aiplatform.edgeDevices.create aiplatform.edgeDevices.delete aiplatform.edgeDevices.get aiplatform.edgeDevices.list aiplatform.edgeDevices.update aiplatform.endpoints.create aiplatform.endpoints.delete aiplatform.endpoints.deploy aiplatform.endpoints.explain aiplatform.endpoints.get aiplatform.endpoints.list aiplatform.endpoints.predict aiplatform.endpoints.undeploy aiplatform.endpoints.update aiplatform.entityTypes.create aiplatform.entityTypes.delete aiplatform.entityTypes.deleteFeatureValues aiplatform.entityTypes.exportFeatureValues aiplatform.entityTypes.get aiplatform.entityTypes.getIamPolicy aiplatform.entityTypes.importFeatureValues aiplatform.entityTypes.list aiplatform.entityTypes.readFeatureValues aiplatform.entityTypes.setIamPolicy aiplatform.entityTypes.streamingReadFeatureValues aiplatform.entityTypes.update aiplatform.entityTypes.writeFeatureValues aiplatform.executions.addExecutionEvents aiplatform.executions.create aiplatform.executions.delete aiplatform.executions.get aiplatform.executions.list aiplatform.executions.queryExecutionInputsAndOutputs aiplatform.executions.update aiplatform.features.create aiplatform.features.delete aiplatform.features.get aiplatform.features.list aiplatform.features.update aiplatform.featurestores.batchReadFeatureValues aiplatform.featurestores.create aiplatform.featurestores.delete aiplatform.featurestores.exportFeatures aiplatform.featurestores.get aiplatform.featurestores.getIamPolicy aiplatform.featurestores.importFeatures aiplatform.featurestores.list aiplatform.featurestores.readFeatures aiplatform.featurestores.setIamPolicy aiplatform.featurestores.update aiplatform.featurestores.writeFeatures aiplatform.humanInTheLoops.cancel aiplatform.humanInTheLoops.create aiplatform.humanInTheLoops.delete aiplatform.humanInTheLoops.get aiplatform.humanInTheLoops.list aiplatform.humanInTheLoops.queryAnnotationStats aiplatform.humanInTheLoops.send aiplatform.humanInTheLoops.update aiplatform.hyperparameterTuningJobs.cancel aiplatform.hyperparameterTuningJobs.create aiplatform.hyperparameterTuningJobs.delete aiplatform.hyperparameterTuningJobs.get aiplatform.hyperparameterTuningJobs.list aiplatform.indexEndpoints.create aiplatform.indexEndpoints.delete aiplatform.indexEndpoints.deploy aiplatform.indexEndpoints.get aiplatform.indexEndpoints.list aiplatform.indexEndpoints.undeploy aiplatform.indexEndpoints.update aiplatform.indexes.create aiplatform.indexes.delete aiplatform.indexes.get aiplatform.indexes.list aiplatform.indexes.update aiplatform.locations.get aiplatform.locations.list aiplatform.metadataSchemas.create aiplatform.metadataSchemas.delete aiplatform.metadataSchemas.get aiplatform.metadataSchemas.list aiplatform.metadataStores.create aiplatform.metadataStores.delete aiplatform.metadataStores.get aiplatform.metadataStores.list aiplatform.migratableResources.migrate aiplatform.migratableResources.search aiplatform.modelDeploymentMonitoringJobs.create aiplatform.modelDeploymentMonitoringJobs.delete aiplatform.modelDeploymentMonitoringJobs.get aiplatform.modelDeploymentMonitoringJobs.list aiplatform.modelDeploymentMonitoringJobs.pause aiplatform.modelDeploymentMonitoringJobs.resume aiplatform.modelDeploymentMonitoringJobs.searchStatsAnomalies aiplatform.modelDeploymentMonitoringJobs.update aiplatform.modelEvaluationSlices.get aiplatform.modelEvaluationSlices.list aiplatform.modelEvaluations.exportEvaluatedDataItems aiplatform.modelEvaluations.get aiplatform.modelEvaluations.list aiplatform.models.delete aiplatform.models.export aiplatform.models.get aiplatform.models.list aiplatform.models.update aiplatform.models.upload aiplatform.nasJobs.cancel aiplatform.nasJobs.create aiplatform.nasJobs.delete aiplatform.nasJobs.get aiplatform.nasJobs.list aiplatform.nasTrialDetails.get aiplatform.nasTrialDetails.list aiplatform.operations.list aiplatform.pipelineJobs.cancel aiplatform.pipelineJobs.create aiplatform.pipelineJobs.delete aiplatform.pipelineJobs.get aiplatform.pipelineJobs.list aiplatform.specialistPools.create aiplatform.specialistPools.delete aiplatform.specialistPools.get aiplatform.specialistPools.list aiplatform.specialistPools.update aiplatform.studies.create aiplatform.studies.delete aiplatform.studies.get aiplatform.studies.list aiplatform.studies.update aiplatform.tensorboardExperiments.create aiplatform.tensorboardExperiments.delete aiplatform.tensorboardExperiments.get aiplatform.tensorboardExperiments.list aiplatform.tensorboardExperiments.update aiplatform.tensorboardExperiments.write aiplatform.tensorboardRuns.batchCreate aiplatform.tensorboardRuns.create aiplatform.tensorboardRuns.delete aiplatform.tensorboardRuns.get aiplatform.tensorboardRuns.list aiplatform.tensorboardRuns.update aiplatform.tensorboardRuns.write aiplatform.tensorboardTimeSeries.batchCreate aiplatform.tensorboardTimeSeries.batchRead aiplatform.tensorboardTimeSeries.create aiplatform.tensorboardTimeSeries.delete aiplatform.tensorboardTimeSeries.get aiplatform.tensorboardTimeSeries.list aiplatform.tensorboardTimeSeries.read aiplatform.tensorboardTimeSeries.update aiplatform.tensorboards.create aiplatform.tensorboards.delete aiplatform.tensorboards.get aiplatform.tensorboards.list aiplatform.tensorboards.recordAccess aiplatform.tensorboards.update aiplatform.trainingPipelines.cancel aiplatform.trainingPipelines.create aiplatform.trainingPipelines.delete aiplatform.trainingPipelines.get aiplatform.trainingPipelines.list aiplatform.trials.create aiplatform.trials.delete aiplatform.trials.get aiplatform.trials.list aiplatform.trials.update resourcemanager.projects.get resourcemanager.projects.list
Proprietário de EntityType da Vertex AI Feature Store (roles/aiplatform.entityTypeOwner) Dá acesso total a todas as permissões para um recurso específico de um tipo de entidade. Recursos de nível mais baixo em que você pode conceder esse papel: Tipo de entidade Contém 1 permissão de proprietário	aiplatform.entityTypes.delete aiplatform.entityTypes.deleteFeatureValues aiplatform.entityTypes.exportFeatureValues aiplatform.entityTypes.get aiplatform.entityTypes.getIamPolicy aiplatform.entityTypes.importFeatureValues aiplatform.entityTypes.readFeatureValues aiplatform.entityTypes.setIamPolicy aiplatform.entityTypes.streamingReadFeatureValues aiplatform.entityTypes.update aiplatform.entityTypes.writeFeatureValues aiplatform.features.* aiplatform.features.create aiplatform.features.delete aiplatform.features.get aiplatform.features.list aiplatform.features.update aiplatform.featurestores.batchReadFeatureValues resourcemanager.projects.get resourcemanager.projects.list
Administrador da Vertex AI Feature Store (roles/aiplatform.featurestoreAdmin) Concede acesso total a todos os recursos na Vertex AI Feature Store Recursos de nível mais baixo em que você pode conceder esse papel: Tipo de entidade Contém 2 permissões de proprietário	aiplatform.entityTypes.* aiplatform.entityTypes.create aiplatform.entityTypes.delete aiplatform.entityTypes.deleteFeatureValues aiplatform.entityTypes.exportFeatureValues aiplatform.entityTypes.get aiplatform.entityTypes.getIamPolicy aiplatform.entityTypes.importFeatureValues aiplatform.entityTypes.list aiplatform.entityTypes.readFeatureValues aiplatform.entityTypes.setIamPolicy aiplatform.entityTypes.streamingReadFeatureValues aiplatform.entityTypes.update aiplatform.entityTypes.writeFeatureValues aiplatform.features.* aiplatform.features.create aiplatform.features.delete aiplatform.features.get aiplatform.features.list aiplatform.features.update aiplatform.featurestores.* aiplatform.featurestores.batchReadFeatureValues aiplatform.featurestores.create aiplatform.featurestores.delete aiplatform.featurestores.exportFeatures aiplatform.featurestores.get aiplatform.featurestores.getIamPolicy aiplatform.featurestores.importFeatures aiplatform.featurestores.list aiplatform.featurestores.readFeatures aiplatform.featurestores.setIamPolicy aiplatform.featurestores.update aiplatform.featurestores.writeFeatures aiplatform.operations.list resourcemanager.projects.get resourcemanager.projects.list
Leitor de dados do Feature Store do Vertex AI (roles/aiplatform.featurestoreDataViewer) Este papel fornece permissões para ler dados de atributos. Recursos de nível mais baixo em que você pode conceder esse papel: Tipo de entidade	aiplatform.entityTypes.exportFeatureValues aiplatform.entityTypes.get aiplatform.entityTypes.readFeatureValues aiplatform.entityTypes.streamingReadFeatureValues aiplatform.features.get aiplatform.features.list aiplatform.featurestores.batchReadFeatureValues resourcemanager.projects.get resourcemanager.projects.list
Gravador de dados do Feature Store do Vertex AI (roles/aiplatform.featurestoreDataWriter) Este papel fornece permissões para ler e gravar dados de recursos. Recursos de nível mais baixo em que você pode conceder esse papel: Tipo de entidade	aiplatform.entityTypes.deleteFeatureValues aiplatform.entityTypes.exportFeatureValues aiplatform.entityTypes.get aiplatform.entityTypes.importFeatureValues aiplatform.entityTypes.readFeatureValues aiplatform.entityTypes.streamingReadFeatureValues aiplatform.entityTypes.writeFeatureValues aiplatform.features.get aiplatform.features.list aiplatform.featurestores.batchReadFeatureValues resourcemanager.projects.get resourcemanager.projects.list
Criador de instância do Feature Store do Vertex AI (roles/aiplatform.featurestoreInstanceCreator) Administrador de recursos do Featurestore, mas não os recursos filhos do Featurestores. Recursos de nível mais baixo em que você pode conceder esse papel: Featurestore	aiplatform.featurestores.create aiplatform.featurestores.delete aiplatform.featurestores.get aiplatform.featurestores.list aiplatform.featurestores.update
Leitor de recursos do Feature Store do Vertex AI (roles/aiplatform.featurestoreResourceViewer) Leitor de todos os recursos no Feature Store do Vertex AI, mas não pode fazer alterações. Recursos de nível mais baixo em que você pode conceder esse papel: Tipo de entidade	aiplatform.entityTypes.get aiplatform.entityTypes.list aiplatform.features.get aiplatform.features.list aiplatform.featurestores.get aiplatform.featurestores.list aiplatform.operations.list resourcemanager.projects.get resourcemanager.projects.list
Usuário da Feature Store da Vertex AI ^Beta (roles/aiplatform.featurestoreUser) Obsoleto. Em vez dele, use o featurestoreAdmin. Contém 2 permissões de proprietário	aiplatform.entityTypes.* aiplatform.entityTypes.create aiplatform.entityTypes.delete aiplatform.entityTypes.deleteFeatureValues aiplatform.entityTypes.exportFeatureValues aiplatform.entityTypes.get aiplatform.entityTypes.getIamPolicy aiplatform.entityTypes.importFeatureValues aiplatform.entityTypes.list aiplatform.entityTypes.readFeatureValues aiplatform.entityTypes.setIamPolicy aiplatform.entityTypes.streamingReadFeatureValues aiplatform.entityTypes.update aiplatform.entityTypes.writeFeatureValues aiplatform.features.* aiplatform.features.create aiplatform.features.delete aiplatform.features.get aiplatform.features.list aiplatform.features.update aiplatform.featurestores.* aiplatform.featurestores.batchReadFeatureValues aiplatform.featurestores.create aiplatform.featurestores.delete aiplatform.featurestores.exportFeatures aiplatform.featurestores.get aiplatform.featurestores.getIamPolicy aiplatform.featurestores.importFeatures aiplatform.featurestores.list aiplatform.featurestores.readFeatures aiplatform.featurestores.setIamPolicy aiplatform.featurestores.update aiplatform.featurestores.writeFeatures aiplatform.operations.list resourcemanager.projects.get resourcemanager.projects.list
Usuário do serviço de migração da Vertex AI (roles/aiplatform.migrator) Concede acesso ao serviço de migração na Vertex AI	aiplatform.migratableResources.* aiplatform.migratableResources.migrate aiplatform.migratableResources.search
Usuário do app TensorBoard da Vertex AI para Web ^Beta (roles/aiplatform.tensorboardWebAppUser) Fornece acesso ao app TensorBoard da Vertex AI para Web. O uso desse app é cobrado. Contém 1 permissão de proprietário	aiplatform.tensorboards.recordAccess
Usuário da Vertex AI (roles/aiplatform.user) Concede acesso para usar todos os recursos na Vertex AI	aiplatform.annotationSpecs.* aiplatform.annotationSpecs.create aiplatform.annotationSpecs.delete aiplatform.annotationSpecs.get aiplatform.annotationSpecs.list aiplatform.annotationSpecs.update aiplatform.annotations.* aiplatform.annotations.create aiplatform.annotations.delete aiplatform.annotations.get aiplatform.annotations.list aiplatform.annotations.update aiplatform.artifacts.* aiplatform.artifacts.create aiplatform.artifacts.delete aiplatform.artifacts.get aiplatform.artifacts.list aiplatform.artifacts.update aiplatform.batchPredictionJobs.* aiplatform.batchPredictionJobs.cancel aiplatform.batchPredictionJobs.create aiplatform.batchPredictionJobs.delete aiplatform.batchPredictionJobs.get aiplatform.batchPredictionJobs.list aiplatform.contexts.* aiplatform.contexts.addContextArtifactsAndExecutions aiplatform.contexts.addContextChildren aiplatform.contexts.create aiplatform.contexts.delete aiplatform.contexts.get aiplatform.contexts.list aiplatform.contexts.queryContextLineageSubgraph aiplatform.contexts.update aiplatform.customJobs.* aiplatform.customJobs.cancel aiplatform.customJobs.create aiplatform.customJobs.delete aiplatform.customJobs.get aiplatform.customJobs.list aiplatform.dataItems.* aiplatform.dataItems.create aiplatform.dataItems.delete aiplatform.dataItems.get aiplatform.dataItems.list aiplatform.dataItems.update aiplatform.dataLabelingJobs.* aiplatform.dataLabelingJobs.cancel aiplatform.dataLabelingJobs.create aiplatform.dataLabelingJobs.delete aiplatform.dataLabelingJobs.get aiplatform.dataLabelingJobs.list aiplatform.datasets.* aiplatform.datasets.create aiplatform.datasets.delete aiplatform.datasets.export aiplatform.datasets.get aiplatform.datasets.import aiplatform.datasets.list aiplatform.datasets.update aiplatform.deploymentResourcePools.* aiplatform.deploymentResourcePools.create aiplatform.deploymentResourcePools.delete aiplatform.deploymentResourcePools.get aiplatform.deploymentResourcePools.list aiplatform.deploymentResourcePools.queryDeployedModels aiplatform.deploymentResourcePools.update aiplatform.edgeDeploymentJobs.* aiplatform.edgeDeploymentJobs.create aiplatform.edgeDeploymentJobs.delete aiplatform.edgeDeploymentJobs.get aiplatform.edgeDeploymentJobs.list aiplatform.edgeDeviceDebugInfo.get aiplatform.edgeDevices.* aiplatform.edgeDevices.create aiplatform.edgeDevices.delete aiplatform.edgeDevices.get aiplatform.edgeDevices.list aiplatform.edgeDevices.update aiplatform.endpoints.* aiplatform.endpoints.create aiplatform.endpoints.delete aiplatform.endpoints.deploy aiplatform.endpoints.explain aiplatform.endpoints.get aiplatform.endpoints.list aiplatform.endpoints.predict aiplatform.endpoints.undeploy aiplatform.endpoints.update aiplatform.entityTypes.create aiplatform.entityTypes.delete aiplatform.entityTypes.deleteFeatureValues aiplatform.entityTypes.exportFeatureValues aiplatform.entityTypes.get aiplatform.entityTypes.importFeatureValues aiplatform.entityTypes.list aiplatform.entityTypes.readFeatureValues aiplatform.entityTypes.streamingReadFeatureValues aiplatform.entityTypes.update aiplatform.entityTypes.writeFeatureValues aiplatform.executions.* aiplatform.executions.addExecutionEvents aiplatform.executions.create aiplatform.executions.delete aiplatform.executions.get aiplatform.executions.list aiplatform.executions.queryExecutionInputsAndOutputs aiplatform.executions.update aiplatform.features.* aiplatform.features.create aiplatform.features.delete aiplatform.features.get aiplatform.features.list aiplatform.features.update aiplatform.featurestores.batchReadFeatureValues aiplatform.featurestores.create aiplatform.featurestores.delete aiplatform.featurestores.exportFeatures aiplatform.featurestores.get aiplatform.featurestores.importFeatures aiplatform.featurestores.list aiplatform.featurestores.readFeatures aiplatform.featurestores.update aiplatform.featurestores.writeFeatures aiplatform.humanInTheLoops.* aiplatform.humanInTheLoops.cancel aiplatform.humanInTheLoops.create aiplatform.humanInTheLoops.delete aiplatform.humanInTheLoops.get aiplatform.humanInTheLoops.list aiplatform.humanInTheLoops.queryAnnotationStats aiplatform.humanInTheLoops.send aiplatform.humanInTheLoops.update aiplatform.hyperparameterTuningJobs.* aiplatform.hyperparameterTuningJobs.cancel aiplatform.hyperparameterTuningJobs.create aiplatform.hyperparameterTuningJobs.delete aiplatform.hyperparameterTuningJobs.get aiplatform.hyperparameterTuningJobs.list aiplatform.indexEndpoints.* aiplatform.indexEndpoints.create aiplatform.indexEndpoints.delete aiplatform.indexEndpoints.deploy aiplatform.indexEndpoints.get aiplatform.indexEndpoints.list aiplatform.indexEndpoints.undeploy aiplatform.indexEndpoints.update aiplatform.indexes.* aiplatform.indexes.create aiplatform.indexes.delete aiplatform.indexes.get aiplatform.indexes.list aiplatform.indexes.update aiplatform.locations.* aiplatform.locations.get aiplatform.locations.list aiplatform.metadataSchemas.* aiplatform.metadataSchemas.create aiplatform.metadataSchemas.delete aiplatform.metadataSchemas.get aiplatform.metadataSchemas.list aiplatform.metadataStores.* aiplatform.metadataStores.create aiplatform.metadataStores.delete aiplatform.metadataStores.get aiplatform.metadataStores.list aiplatform.modelDeploymentMonitoringJobs.* aiplatform.modelDeploymentMonitoringJobs.create aiplatform.modelDeploymentMonitoringJobs.delete aiplatform.modelDeploymentMonitoringJobs.get aiplatform.modelDeploymentMonitoringJobs.list aiplatform.modelDeploymentMonitoringJobs.pause aiplatform.modelDeploymentMonitoringJobs.resume aiplatform.modelDeploymentMonitoringJobs.searchStatsAnomalies aiplatform.modelDeploymentMonitoringJobs.update aiplatform.modelEvaluationSlices.* aiplatform.modelEvaluationSlices.get aiplatform.modelEvaluationSlices.list aiplatform.modelEvaluations.* aiplatform.modelEvaluations.exportEvaluatedDataItems aiplatform.modelEvaluations.get aiplatform.modelEvaluations.list aiplatform.models.* aiplatform.models.delete aiplatform.models.export aiplatform.models.get aiplatform.models.list aiplatform.models.update aiplatform.models.upload aiplatform.nasJobs.* aiplatform.nasJobs.cancel aiplatform.nasJobs.create aiplatform.nasJobs.delete aiplatform.nasJobs.get aiplatform.nasJobs.list aiplatform.nasTrialDetails.* aiplatform.nasTrialDetails.get aiplatform.nasTrialDetails.list aiplatform.operations.list aiplatform.pipelineJobs.* aiplatform.pipelineJobs.cancel aiplatform.pipelineJobs.create aiplatform.pipelineJobs.delete aiplatform.pipelineJobs.get aiplatform.pipelineJobs.list aiplatform.specialistPools.* aiplatform.specialistPools.create aiplatform.specialistPools.delete aiplatform.specialistPools.get aiplatform.specialistPools.list aiplatform.specialistPools.update aiplatform.studies.* aiplatform.studies.create aiplatform.studies.delete aiplatform.studies.get aiplatform.studies.list aiplatform.studies.update aiplatform.tensorboardExperiments.* aiplatform.tensorboardExperiments.create aiplatform.tensorboardExperiments.delete aiplatform.tensorboardExperiments.get aiplatform.tensorboardExperiments.list aiplatform.tensorboardExperiments.update aiplatform.tensorboardExperiments.write aiplatform.tensorboardRuns.* aiplatform.tensorboardRuns.batchCreate aiplatform.tensorboardRuns.create aiplatform.tensorboardRuns.delete aiplatform.tensorboardRuns.get aiplatform.tensorboardRuns.list aiplatform.tensorboardRuns.update aiplatform.tensorboardRuns.write aiplatform.tensorboardTimeSeries.* aiplatform.tensorboardTimeSeries.batchCreate aiplatform.tensorboardTimeSeries.batchRead aiplatform.tensorboardTimeSeries.create aiplatform.tensorboardTimeSeries.delete aiplatform.tensorboardTimeSeries.get aiplatform.tensorboardTimeSeries.list aiplatform.tensorboardTimeSeries.read aiplatform.tensorboardTimeSeries.update aiplatform.tensorboards.create aiplatform.tensorboards.delete aiplatform.tensorboards.get aiplatform.tensorboards.list aiplatform.tensorboards.update aiplatform.trainingPipelines.* aiplatform.trainingPipelines.cancel aiplatform.trainingPipelines.create aiplatform.trainingPipelines.delete aiplatform.trainingPipelines.get aiplatform.trainingPipelines.list aiplatform.trials.* aiplatform.trials.create aiplatform.trials.delete aiplatform.trials.get aiplatform.trials.list aiplatform.trials.update resourcemanager.projects.get resourcemanager.projects.list
Leitor da Vertex AI (roles/aiplatform.viewer) Concede acesso para visualizar todos os recursos na Vertex AI	aiplatform.annotationSpecs.get aiplatform.annotationSpecs.list aiplatform.annotations.get aiplatform.annotations.list aiplatform.artifacts.get aiplatform.artifacts.list aiplatform.batchPredictionJobs.get aiplatform.batchPredictionJobs.list aiplatform.contexts.get aiplatform.contexts.list aiplatform.contexts.queryContextLineageSubgraph aiplatform.customJobs.get aiplatform.customJobs.list aiplatform.dataItems.get aiplatform.dataItems.list aiplatform.dataLabelingJobs.get aiplatform.dataLabelingJobs.list aiplatform.datasets.get aiplatform.datasets.list aiplatform.deploymentResourcePools.get aiplatform.deploymentResourcePools.list aiplatform.deploymentResourcePools.queryDeployedModels aiplatform.edgeDeploymentJobs.get aiplatform.edgeDeploymentJobs.list aiplatform.edgeDeviceDebugInfo.get aiplatform.edgeDevices.get aiplatform.edgeDevices.list aiplatform.endpoints.get aiplatform.endpoints.list aiplatform.entityTypes.get aiplatform.entityTypes.list aiplatform.executions.get aiplatform.executions.list aiplatform.executions.queryExecutionInputsAndOutputs aiplatform.features.get aiplatform.features.list aiplatform.featurestores.get aiplatform.featurestores.list aiplatform.humanInTheLoops.get aiplatform.humanInTheLoops.list aiplatform.hyperparameterTuningJobs.get aiplatform.hyperparameterTuningJobs.list aiplatform.indexEndpoints.get aiplatform.indexEndpoints.list aiplatform.indexes.get aiplatform.indexes.list aiplatform.locations.* aiplatform.locations.get aiplatform.locations.list aiplatform.metadataSchemas.get aiplatform.metadataSchemas.list aiplatform.metadataStores.get aiplatform.metadataStores.list aiplatform.modelDeploymentMonitoringJobs.get aiplatform.modelDeploymentMonitoringJobs.list aiplatform.modelDeploymentMonitoringJobs.searchStatsAnomalies aiplatform.modelEvaluationSlices.* aiplatform.modelEvaluationSlices.get aiplatform.modelEvaluationSlices.list aiplatform.modelEvaluations.get aiplatform.modelEvaluations.list aiplatform.models.get aiplatform.models.list aiplatform.nasJobs.get aiplatform.nasJobs.list aiplatform.nasTrialDetails.* aiplatform.nasTrialDetails.get aiplatform.nasTrialDetails.list aiplatform.operations.list aiplatform.pipelineJobs.get aiplatform.pipelineJobs.list aiplatform.specialistPools.get aiplatform.specialistPools.list aiplatform.specialistPools.update aiplatform.studies.get aiplatform.studies.list aiplatform.tensorboardExperiments.get aiplatform.tensorboardExperiments.list aiplatform.tensorboardRuns.get aiplatform.tensorboardRuns.list aiplatform.tensorboardTimeSeries.batchRead aiplatform.tensorboardTimeSeries.get aiplatform.tensorboardTimeSeries.list aiplatform.tensorboardTimeSeries.read aiplatform.tensorboards.get aiplatform.tensorboards.list aiplatform.trainingPipelines.get aiplatform.trainingPipelines.list aiplatform.trials.get aiplatform.trials.list resourcemanager.projects.get resourcemanager.projects.list

Papéis básicos

Os papéis básicos mais antigos do Google Cloud são comuns a todos os serviços do Google Cloud. Esses papéis são de Proprietário, Editor e Visualizador.

Os papéis básicos fornecem permissões em todo o Google Cloud, não apenas na Vertex AI. Por esse motivo, use os papéis da Vertex AI sempre que possível.

Papéis personalizados

Se os papéis predefinidos do IAM para a Vertex AI não atenderem às suas necessidades, é possível definir papéis personalizados. Os papéis personalizados permitem escolher um conjunto específico de permissões, criar seu próprio papel com elas e concedê-lo aos usuários da organização. Para mais informações, consulte Noções básicas sobre papéis personalizados do IAM.

Políticas no nível do projeto e no nível do recurso

Definir uma política no nível do recurso não afeta as políticas no nível do projeto. Um recurso herda todas as políticas do ancestral. Você pode usar esses dois níveis de granularidade para personalizar permissões. Por exemplo, é possível conceder aos usuários permissões de leitura no nível do projeto para que eles leiam todos os recursos no projeto. Em seguida, conceda aos usuários permissões de gravação por recurso (no nível do recurso).

Nem todos os papéis e recursos predefinidos da Vertex AI são compatíveis com políticas no nível do recurso. Para ver quais papéis podem ser usados em quais recursos, veja as descrições de cada um.

Recursos compatíveis

A Vertex AI é compatível com a featurestore e os recursos de tipo de entidade da Vertex AI Feature Store. Para mais informações, consulte Controlar o acesso aos recursos da Vertex AI Feature Store.

Depois de conceder ou revogar o acesso a um recurso, essas alterações levam algum tempo para serem propagadas. Para mais informações, consulte as Perguntas frequentes do IAM.

Sobre contas de serviço e agentes de serviço

Contas de serviço

Uma conta de serviço é um tipo especial de conta usada por um aplicativo ou instância de máquina virtual (VM), não uma pessoa. É possível criar e atribuir permissões a contas de serviço para fornecer permissões específicas a um recurso ou aplicativo.

Para informações sobre como usar uma conta de serviço para personalizar as permissões disponíveis para um contêiner de treinamento personalizado ou um contêiner que disponibilize predições on-line para um modelo treinado personalizado, leiaComo usar uma conta de serviço personalizada para criar um anexo da VLAN de monitoramento.

As contas de serviço são identificadas por um endereço de e-mail.

Agentes de serviço

Os agentes de serviço são contas de serviço gerenciadas pelo Google que são fornecidas automaticamente. eles permitem que um serviço acesse recursos em seu nome. A Vertex AI usa estes agentes de serviço:

Nome	Usado para	Endereço de e-mail
Agente de serviço da Vertex AI	Funcionalidade da Vertex AI	[e-mail]
Agente de serviço de código personalizado da Vertex AI	Código de treinamento personalizado	[e-mail]

O Agente de serviço de código personalizado da Vertex AI será criado somente se você executar um código de treinamento personalizado para treinar um modelo personalizado.

Quando criado, cada agente de serviço recebe um dos seguintes papéis predefinidos do projeto. Cada agente de serviço recebe o papel que corresponde ao nome.

Papel	Permissões
Agente de serviço da Vertex AI (roles/aiplatform.serviceAgent) Concede à Vertex AI as permissões necessárias para funcionar. Contém 1 permissão de proprietário	aiplatform.annotationSpecs.* aiplatform.annotationSpecs.create aiplatform.annotationSpecs.delete aiplatform.annotationSpecs.get aiplatform.annotationSpecs.list aiplatform.annotationSpecs.update aiplatform.annotations.* aiplatform.annotations.create aiplatform.annotations.delete aiplatform.annotations.get aiplatform.annotations.list aiplatform.annotations.update aiplatform.artifacts.* aiplatform.artifacts.create aiplatform.artifacts.delete aiplatform.artifacts.get aiplatform.artifacts.list aiplatform.artifacts.update aiplatform.batchPredictionJobs.* aiplatform.batchPredictionJobs.cancel aiplatform.batchPredictionJobs.create aiplatform.batchPredictionJobs.delete aiplatform.batchPredictionJobs.get aiplatform.batchPredictionJobs.list aiplatform.contexts.* aiplatform.contexts.addContextArtifactsAndExecutions aiplatform.contexts.addContextChildren aiplatform.contexts.create aiplatform.contexts.delete aiplatform.contexts.get aiplatform.contexts.list aiplatform.contexts.queryContextLineageSubgraph aiplatform.contexts.update aiplatform.customJobs.* aiplatform.customJobs.cancel aiplatform.customJobs.create aiplatform.customJobs.delete aiplatform.customJobs.get aiplatform.customJobs.list aiplatform.dataItems.* aiplatform.dataItems.create aiplatform.dataItems.delete aiplatform.dataItems.get aiplatform.dataItems.list aiplatform.dataItems.update aiplatform.dataLabelingJobs.* aiplatform.dataLabelingJobs.cancel aiplatform.dataLabelingJobs.create aiplatform.dataLabelingJobs.delete aiplatform.dataLabelingJobs.get aiplatform.dataLabelingJobs.list aiplatform.datasets.* aiplatform.datasets.create aiplatform.datasets.delete aiplatform.datasets.export aiplatform.datasets.get aiplatform.datasets.import aiplatform.datasets.list aiplatform.datasets.update aiplatform.deploymentResourcePools.* aiplatform.deploymentResourcePools.create aiplatform.deploymentResourcePools.delete aiplatform.deploymentResourcePools.get aiplatform.deploymentResourcePools.list aiplatform.deploymentResourcePools.queryDeployedModels aiplatform.deploymentResourcePools.update aiplatform.edgeDeploymentJobs.* aiplatform.edgeDeploymentJobs.create aiplatform.edgeDeploymentJobs.delete aiplatform.edgeDeploymentJobs.get aiplatform.edgeDeploymentJobs.list aiplatform.edgeDeviceDebugInfo.get aiplatform.edgeDevices.* aiplatform.edgeDevices.create aiplatform.edgeDevices.delete aiplatform.edgeDevices.get aiplatform.edgeDevices.list aiplatform.edgeDevices.update aiplatform.endpoints.* aiplatform.endpoints.create aiplatform.endpoints.delete aiplatform.endpoints.deploy aiplatform.endpoints.explain aiplatform.endpoints.get aiplatform.endpoints.list aiplatform.endpoints.predict aiplatform.endpoints.undeploy aiplatform.endpoints.update aiplatform.entityTypes.create aiplatform.entityTypes.delete aiplatform.entityTypes.deleteFeatureValues aiplatform.entityTypes.exportFeatureValues aiplatform.entityTypes.get aiplatform.entityTypes.importFeatureValues aiplatform.entityTypes.list aiplatform.entityTypes.readFeatureValues aiplatform.entityTypes.streamingReadFeatureValues aiplatform.entityTypes.update aiplatform.entityTypes.writeFeatureValues aiplatform.executions.* aiplatform.executions.addExecutionEvents aiplatform.executions.create aiplatform.executions.delete aiplatform.executions.get aiplatform.executions.list aiplatform.executions.queryExecutionInputsAndOutputs aiplatform.executions.update aiplatform.features.* aiplatform.features.create aiplatform.features.delete aiplatform.features.get aiplatform.features.list aiplatform.features.update aiplatform.featurestores.batchReadFeatureValues aiplatform.featurestores.create aiplatform.featurestores.delete aiplatform.featurestores.exportFeatures aiplatform.featurestores.get aiplatform.featurestores.importFeatures aiplatform.featurestores.list aiplatform.featurestores.readFeatures aiplatform.featurestores.update aiplatform.featurestores.writeFeatures aiplatform.humanInTheLoops.* aiplatform.humanInTheLoops.create aiplatform.humanInTheLoops.delete aiplatform.humanInTheLoops.get aiplatform.humanInTheLoops.list aiplatform.humanInTheLoops.queryAnnotationStats aiplatform.humanInTheLoops.send aiplatform.humanInTheLoops.update aiplatform.hyperparameterTuningJobs.* aiplatform.hyperparameterTuningJobs.cancel aiplatform.hyperparameterTuningJobs.create aiplatform.hyperparameterTuningJobs.delete aiplatform.hyperparameterTuningJobs.get aiplatform.hyperparameterTuningJobs.list aiplatform.indexEndpoints.* aiplatform.indexEndpoints.create aiplatform.indexEndpoints.delete aiplatform.indexEndpoints.deploy aiplatform.indexEndpoints.get aiplatform.indexEndpoints.list aiplatform.indexEndpoints.undeploy aiplatform.indexEndpoints.update aiplatform.indexes.* aiplatform.indexes.create aiplatform.indexes.delete aiplatform.indexes.get aiplatform.indexes.list aiplatform.indexes.update aiplatform.locations.* aiplatform.locations.get aiplatform.locations.list aiplatform.metadataSchemas.* aiplatform.metadataSchemas.create aiplatform.metadataSchemas.delete aiplatform.metadataSchemas.get aiplatform.metadataSchemas.list aiplatform.metadataStores.* aiplatform.metadataStores.create aiplatform.metadataStores.delete aiplatform.metadataStores.get aiplatform.metadataStores.list aiplatform.modelDeploymentMonitoringJobs.* aiplatform.modelDeploymentMonitoringJobs.create aiplatform.modelDeploymentMonitoringJobs.delete aiplatform.modelDeploymentMonitoringJobs.get aiplatform.modelDeploymentMonitoringJobs.list aiplatform.modelDeploymentMonitoringJobs.pause aiplatform.modelDeploymentMonitoringJobs.resume aiplatform.modelDeploymentMonitoringJobs.searchStatsAnomalies aiplatform.modelDeploymentMonitoringJobs.update aiplatform.modelEvaluationSlices.* aiplatform.modelEvaluationSlices.get aiplatform.modelEvaluationSlices.list aiplatform.modelEvaluations.* aiplatform.modelEvaluations.exportEvaluatedDataItems aiplatform.modelEvaluations.get aiplatform.modelEvaluations.list aiplatform.models.* aiplatform.models.delete aiplatform.models.export aiplatform.models.get aiplatform.models.list aiplatform.models.update aiplatform.models.upload aiplatform.nasJobs.* aiplatform.nasJobs.cancel aiplatform.nasJobs.create aiplatform.nasJobs.delete aiplatform.nasJobs.get aiplatform.nasJobs.list aiplatform.nasTrialDetails.* aiplatform.nasTrialDetails.get aiplatform.nasTrialDetails.list aiplatform.operations.list aiplatform.pipelineJobs.* aiplatform.pipelineJobs.cancel aiplatform.pipelineJobs.create aiplatform.pipelineJobs.delete aiplatform.pipelineJobs.get aiplatform.pipelineJobs.list aiplatform.specialistPools.* aiplatform.specialistPools.create aiplatform.specialistPools.delete aiplatform.specialistPools.get aiplatform.specialistPools.list aiplatform.specialistPools.update aiplatform.studies.* aiplatform.studies.create aiplatform.studies.delete aiplatform.studies.get aiplatform.studies.list aiplatform.studies.update aiplatform.tensorboardExperiments.* aiplatform.tensorboardExperiments.create aiplatform.tensorboardExperiments.delete aiplatform.tensorboardExperiments.get aiplatform.tensorboardExperiments.list aiplatform.tensorboardExperiments.update aiplatform.tensorboardExperiments.write aiplatform.tensorboardRuns.* aiplatform.tensorboardRuns.batchCreate aiplatform.tensorboardRuns.create aiplatform.tensorboardRuns.delete aiplatform.tensorboardRuns.get aiplatform.tensorboardRuns.list aiplatform.tensorboardRuns.update aiplatform.tensorboardRuns.write aiplatform.tensorboardTimeSeries.* aiplatform.tensorboardTimeSeries.batchCreate aiplatform.tensorboardTimeSeries.batchRead aiplatform.tensorboardTimeSeries.create aiplatform.tensorboardTimeSeries.delete aiplatform.tensorboardTimeSeries.get aiplatform.tensorboardTimeSeries.list aiplatform.tensorboardTimeSeries.read aiplatform.tensorboardTimeSeries.update aiplatform.tensorboards.create aiplatform.tensorboards.delete aiplatform.tensorboards.get aiplatform.tensorboards.list aiplatform.tensorboards.update aiplatform.trainingPipelines.* aiplatform.trainingPipelines.cancel aiplatform.trainingPipelines.create aiplatform.trainingPipelines.delete aiplatform.trainingPipelines.get aiplatform.trainingPipelines.list aiplatform.trials.* aiplatform.trials.create aiplatform.trials.delete aiplatform.trials.get aiplatform.trials.list aiplatform.trials.update artifactregistry.repositories.create artifactregistry.repositories.downloadArtifacts artifactregistry.repositories.get artifactregistry.repositories.list artifactregistry.repositories.uploadArtifacts artifactregistry.tags.get artifactregistry.versions.get automl.datasets.export automl.datasets.get automl.datasets.list automl.modelEvaluations.list automl.models.get automl.models.list automl.operations.get automl.tableSpecs.get bigquery.datasets.create bigquery.datasets.get bigquery.jobs.create bigquery.jobs.get bigquery.models.create bigquery.models.export bigquery.models.getData bigquery.readsessions.create bigquery.readsessions.getData bigquery.tables.create bigquery.tables.export bigquery.tables.get bigquery.tables.getData bigquery.tables.update bigquery.tables.updateData bigtable.tables.get bigtable.tables.list bigtable.tables.readRows compute.machineTypes.get dataflow.jobs.* dataflow.jobs.cancel dataflow.jobs.create dataflow.jobs.get dataflow.jobs.list dataflow.jobs.snapshot dataflow.jobs.updateContents dataflow.messages.list dataflow.metrics.get dataflow.snapshots.* dataflow.snapshots.delete dataflow.snapshots.get dataflow.snapshots.list datalabeling.annotateddatasets.get datalabeling.datasets.export datalabeling.datasets.get datalabeling.datasets.list datalabeling.operations.get iam.serviceAccounts.actAs iam.serviceAccounts.getAccessToken logging.logEntries.create ml.models.list ml.operations.get ml.versions.get ml.versions.list resourcemanager.projects.get resourcemanager.projects.list serviceusage.services.use storage.buckets.create storage.buckets.delete storage.buckets.get storage.buckets.list storage.objects.create storage.objects.delete storage.objects.get storage.objects.list storage.objects.update
Agente de serviço de código personalizado da Vertex AI (roles/aiplatform.customCodeServiceAgent) Concede as permissões adequadas ao código personalizado da Vertex AI. Contém cinco permissões de proprietário	aiplatform.annotationSpecs.* aiplatform.annotationSpecs.create aiplatform.annotationSpecs.delete aiplatform.annotationSpecs.get aiplatform.annotationSpecs.list aiplatform.annotationSpecs.update aiplatform.annotations.* aiplatform.annotations.create aiplatform.annotations.delete aiplatform.annotations.get aiplatform.annotations.list aiplatform.annotations.update aiplatform.artifacts.* aiplatform.artifacts.create aiplatform.artifacts.delete aiplatform.artifacts.get aiplatform.artifacts.list aiplatform.artifacts.update aiplatform.batchPredictionJobs.* aiplatform.batchPredictionJobs.cancel aiplatform.batchPredictionJobs.create aiplatform.batchPredictionJobs.delete aiplatform.batchPredictionJobs.get aiplatform.batchPredictionJobs.list aiplatform.contexts.* aiplatform.contexts.addContextArtifactsAndExecutions aiplatform.contexts.addContextChildren aiplatform.contexts.create aiplatform.contexts.delete aiplatform.contexts.get aiplatform.contexts.list aiplatform.contexts.queryContextLineageSubgraph aiplatform.contexts.update aiplatform.customJobs.* aiplatform.customJobs.cancel aiplatform.customJobs.create aiplatform.customJobs.delete aiplatform.customJobs.get aiplatform.customJobs.list aiplatform.dataItems.* aiplatform.dataItems.create aiplatform.dataItems.delete aiplatform.dataItems.get aiplatform.dataItems.list aiplatform.dataItems.update aiplatform.dataLabelingJobs.* aiplatform.dataLabelingJobs.cancel aiplatform.dataLabelingJobs.create aiplatform.dataLabelingJobs.delete aiplatform.dataLabelingJobs.get aiplatform.dataLabelingJobs.list aiplatform.datasets.* aiplatform.datasets.create aiplatform.datasets.delete aiplatform.datasets.export aiplatform.datasets.get aiplatform.datasets.import aiplatform.datasets.list aiplatform.datasets.update aiplatform.deploymentResourcePools.* aiplatform.deploymentResourcePools.create aiplatform.deploymentResourcePools.delete aiplatform.deploymentResourcePools.get aiplatform.deploymentResourcePools.list aiplatform.deploymentResourcePools.queryDeployedModels aiplatform.deploymentResourcePools.update aiplatform.edgeDeploymentJobs.* aiplatform.edgeDeploymentJobs.create aiplatform.edgeDeploymentJobs.delete aiplatform.edgeDeploymentJobs.get aiplatform.edgeDeploymentJobs.list aiplatform.edgeDeviceDebugInfo.get aiplatform.edgeDevices.* aiplatform.edgeDevices.create aiplatform.edgeDevices.delete aiplatform.edgeDevices.get aiplatform.edgeDevices.list aiplatform.edgeDevices.update aiplatform.endpoints.* aiplatform.endpoints.create aiplatform.endpoints.delete aiplatform.endpoints.deploy aiplatform.endpoints.explain aiplatform.endpoints.get aiplatform.endpoints.list aiplatform.endpoints.predict aiplatform.endpoints.undeploy aiplatform.endpoints.update aiplatform.entityTypes.create aiplatform.entityTypes.delete aiplatform.entityTypes.deleteFeatureValues aiplatform.entityTypes.exportFeatureValues aiplatform.entityTypes.get aiplatform.entityTypes.importFeatureValues aiplatform.entityTypes.list aiplatform.entityTypes.readFeatureValues aiplatform.entityTypes.streamingReadFeatureValues aiplatform.entityTypes.update aiplatform.entityTypes.writeFeatureValues aiplatform.executions.* aiplatform.executions.addExecutionEvents aiplatform.executions.create aiplatform.executions.delete aiplatform.executions.get aiplatform.executions.list aiplatform.executions.queryExecutionInputsAndOutputs aiplatform.executions.update aiplatform.features.* aiplatform.features.create aiplatform.features.delete aiplatform.features.get aiplatform.features.list aiplatform.features.update aiplatform.featurestores.batchReadFeatureValues aiplatform.featurestores.create aiplatform.featurestores.delete aiplatform.featurestores.exportFeatures aiplatform.featurestores.get aiplatform.featurestores.importFeatures aiplatform.featurestores.list aiplatform.featurestores.readFeatures aiplatform.featurestores.update aiplatform.featurestores.writeFeatures aiplatform.humanInTheLoops.* aiplatform.humanInTheLoops.create aiplatform.humanInTheLoops.delete aiplatform.humanInTheLoops.get aiplatform.humanInTheLoops.list aiplatform.humanInTheLoops.queryAnnotationStats aiplatform.humanInTheLoops.send aiplatform.humanInTheLoops.update aiplatform.hyperparameterTuningJobs.* aiplatform.hyperparameterTuningJobs.cancel aiplatform.hyperparameterTuningJobs.create aiplatform.hyperparameterTuningJobs.delete aiplatform.hyperparameterTuningJobs.get aiplatform.hyperparameterTuningJobs.list aiplatform.indexEndpoints.* aiplatform.indexEndpoints.create aiplatform.indexEndpoints.delete aiplatform.indexEndpoints.deploy aiplatform.indexEndpoints.get aiplatform.indexEndpoints.list aiplatform.indexEndpoints.undeploy aiplatform.indexEndpoints.update aiplatform.indexes.* aiplatform.indexes.create aiplatform.indexes.delete aiplatform.indexes.get aiplatform.indexes.list aiplatform.indexes.update aiplatform.locations.* aiplatform.locations.get aiplatform.locations.list aiplatform.metadataSchemas.* aiplatform.metadataSchemas.create aiplatform.metadataSchemas.delete aiplatform.metadataSchemas.get aiplatform.metadataSchemas.list aiplatform.metadataStores.* aiplatform.metadataStores.create aiplatform.metadataStores.delete aiplatform.metadataStores.get aiplatform.metadataStores.list aiplatform.modelDeploymentMonitoringJobs.* aiplatform.modelDeploymentMonitoringJobs.create aiplatform.modelDeploymentMonitoringJobs.delete aiplatform.modelDeploymentMonitoringJobs.get aiplatform.modelDeployment