Google Cloud Platform provider outpost permissions - List of Google Cloud Platform (GCP) permissions for use during Cortex Cloud onboarding outposts to enable continuous monitoring in your cloud environment. - Administrator Guide - Cortex CLOUD

storage.objects.delete

Delete objects. This is used to remove temporary files, logs, or scan results from the artifact bucket once they have been processed or are no longer needed.

storage.objects.delete

Delete objects. This is used to remove temporary files, logs, or scan results from the artifact bucket once they have been processed or are no longer needed.

roles/compute.admin

Grant full administrative control over Compute Engine resources. This allows the outpost to dynamically provision, manage, and delete Virtual Machines, disks, and networks required for scanning operations, ensuring resources are only running when needed.

roles/iam.serviceAccountTokenCreator

Create OAuth2 access tokens or OpenID Connect ID tokens. This allows the CTS service account to impersonate the Cortex Engine service account, facilitating the secure trust chain between the Cortex Platform and the Outpost.

roles/iam.serviceAccountUser

Act as a service account. This permission enables Cortex to launch Compute instances with specific service account identities, such as the scan runner or proxy, ensuring they have the precise permissions needed for their tasks.

roles/pubsub.subscriber

Consume messages from Pub/Sub subscriptions. This facilitates event-driven communication, allowing the Cortex Engine to receive notifications about bucket events or scanning triggers for the bucket-communication process.

storage.objects.create

Create or upload objects. This allows the scanner to write scan results, logs, or other artifacts to the designated communication or customer-data buckets.

storage.objects.delete

Delete objects. This is used to remove temporary files, logs, or scan results from the artifact bucket once they have been processed or are no longer needed.

storage.objects.list

List objects. This allows the DSPM service to list objects written by GCP managed services to the customer-data bucket, facilitating data discovery.

bigquery.jobs.create

Create an export job. This allows the outpost to transfer data from BigQuery to Google Cloud Storage (GCS). This is necessary to facilitate efficient data scanning and classification by moving data to a temporary staging area for analysis without impacting the live dataset.

cloudkms.cryptoKeys.create

Create a new cryptographic key. This key is used specifically for Bigtable encryption operations. It ensures that data extracted or processed during Bigtable scanning remains encrypted and secure.

cloudkms.cryptoKeys.getIamPolicy

Retrieve the IAM policy for a cryptographic key. This allows the system to verify access permissions for the keys used in Bigtable encryption, ensuring that only authorized entities can use them.

cloudkms.cryptoKeys.setIamPolicy

Set the IAM policy for a cryptographic key. This is required to grant the necessary permissions to the scanner service account so it can use the key for Bigtable encryption operations.

cloudkms.cryptoKeys.update

Modify the properties of a cryptographic key. This is used to manage the configuration of keys used for Bigtable encryption, such as updating labels or rotation schedules.

cloudkms.cryptoKeyVersions.create

Create a new version for a cryptographic key. This is used to rotate keys or generate new key material specifically for Bigtable encryption operations.

cloudkms.cryptoKeyVersions.destroy

Permanently destroy a cryptographic key version. This allows for the secure cleanup of key material that is no longer needed after Bigtable scanning is complete, preventing clutter and reducing the risk of key reuse.

cloudkms.cryptoKeyVersions.get

Retrieve the details and metadata of a cryptographic key version. This provides necessary context about the key being used for Bigtable encryption, such as its state and algorithm, to ensure the correct key is applied.

cloudkms.cryptoKeyVersions.list

List all versions of a cryptographic key. This allows the system to identify available key versions for Bigtable encryption and manage their lifecycle effectively during the scanning process.

cloudkms.cryptoKeyVersions.update

Modify the settings and state of a cryptographic key version. This is used to enable or disable key versions used for Bigtable encryption as needed, ensuring control over key usage.

cloudkms.cryptoKeyVersions.useToDecrypt

Use a cryptographic key version to decrypt data. This permission is essential for the scanner to read and analyze encrypted Bigtable data during the scanning process, enabling the classification of sensitive information.

cloudkms.cryptoKeyVersions.useToEncrypt

Use a cryptographic key version to encrypt data. This ensures that any data written or processed during the Bigtable scan is securely encrypted, maintaining data confidentiality throughout the operation.

cloudkms.keyRings.create

Create a new key ring. This is used to organize and hold the cryptographic keys required for Bigtable encryption in a logical group within the project.

cloudsql.databases.create

Create databases. This is performed as part of an instance restore operation. It allows the creation of a temporary database copy for safe scanning without impacting the live production database.

cloudsql.databases.delete

Delete databases. This is used to clean up temporary databases created for scanning once the analysis is complete, ensuring no residual data remains in the environment.

cloudsql.databases.get

Retrieve database details. This allows the system to verify the configuration and status of the database during the restore and scan operation to ensure it is ready for access.

cloudsql.databases.list

List databases. This provides visibility into the databases present on an instance, which is necessary for identifying targets for restore and scanning.

cloudsql.databases.update

Modify database properties. This is used during the instance restore operation to configure the temporary database correctly for access and scanning.

cloudsql.instances.create

Create a Cloud SQL instance. This is used to create a temporary instance for restoring backups. This enables analysis of the data in an isolated environment without affecting the production instance.

cloudsql.instances.delete

Delete Cloud SQL instances. This is a critical cleanup permission. It allows the removal of the temporary instance created for scanning once the task is finished.

cloudsql.instances.get

Retrieve instance details. This provides necessary metadata about the Cloud SQL instance to ensure it is ready for scanning or restore operations.

cloudsql.instances.list

List Cloud SQL instances. This allows the system to discover instances and verify the presence of the temporary instance created for scanning.

cloudsql.instances.restoreBackup

Restore a Cloud SQL instance from a backup. This is the core action for safe scanning. It allows Cortex to restore data to a new, temporary instance for analysis, leaving the source untouched.

cloudsql.instances.update

Modify Cloud SQL instance properties. This is used to configure the temporary instance, such as adjusting network settings to allow the scanner to connect.

cloudsql.users.list

List all users on a Cloud SQL instance. This is used to manage users on the temporary instance during the restore and scan process.

cloudsql.users.update

Modify user settings. This allows the system to update the password or permissions of the temporary user on the Cloud SQL instance to ensure access.

secretmanager.secrets.create

Create a new secret. This is used to securely store credentials or configuration data required by the scanner, such as temporary database passwords, to connect to scan targets.

secretmanager.secrets.delete

Delete a secret. This allows the system to clean up and remove secrets once they are no longer needed, preventing sensitive data from persisting in the environment.

secretmanager.secrets.get

View secret metadata. This allows the system to verify the configuration and existence of secrets used by the scanner without revealing the payload.

secretmanager.secrets.list

List secrets. This provides visibility into the secrets managed within the project, allowing the system to identify those related to scanning operations.

secretmanager.versions.add

Add a new secret version. This allows the system to update the stored secret value, for example, when rotating temporary credentials for a new scan.

cloudsql.instances.connect

Connect to a Cloud SQL instance. This permission allows the scanner to establish a connection to the database instance to perform data scanning and classification. This runtime permission is needed at the monitored project level, not at the outpost project.

cloudsql.instances.createTagBinding

Apply tags to a Cloud SQL instance. This is used to tag temporary instances created for scanning, facilitating resource tracking, cost allocation, and automated cleanup. This runtime permission is needed at the monitored project level, not at the outpost project.

cloudsql.instances.deleteTagBinding

Remove tags from a Cloud SQL instance. This is used during the lifecycle management of the temporary scanning instance to update or clean up metadata. This runtime permission is needed at the monitored project level, not at the outpost project.

cloudsql.instances.listTagBindings

List tags on a Cloud SQL instance. This is used to verify that the correct tags are applied to the temporary scanning instance for tracking purposes. This runtime permission is needed at the monitored project level, not at the outpost project.

cloudsql.instances.login

Log into a Cloud SQL instance. This permission is required for the scanner to authenticate to the database and access the data for classification. This runtime permission is needed at the monitored project level, not at the outpost project.

cloudsql.instances.restart

Restart a Cloud SQL instance. This is used during the instance restore workflow to apply configurations or recover the temporary instance if needed. This runtime permission is needed at the monitored project level, not at the outpost project.

cloudsql.users.create

Create a new user. This is used to create a temporary service user on the restored Cloud SQL instance, granting the scanner access to the data. This runtime permission is needed at the monitored project level, not at the outpost project.

cloudsql.users.delete

Delete a user. This allows for the removal of the temporary service user from the Cloud SQL instance after the scan is complete. This runtime permission is needed at the monitored project level, not at the outpost project.

cloudsql.users.get

Retrieve user details. This allows the system to verify that the temporary user has been correctly created on the instance and has the appropriate attributes. This runtime permission is needed at the monitored project level, not at the outpost project.

roles/bigtable.admin

Grant full administrative control over Bigtable instances and clusters. This is required to manage the lifecycle of Bigtable resources used during data classification and analysis, including the creation and deletion of temporary backups. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.secrets.update

Update secret metadata. This is used to modify labels or settings on secrets, such as marking them for replication or tracking ownership. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.access

Access secret payloads. This permission allows the scanner to retrieve the actual sensitive value (e.g., a password) stored in a secret version to authenticate against a target. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.destroy

Destroy a secret version. This permanently removes a specific version of a secret payload, ensuring that old credentials cannot be recovered. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.disable

Disable a secret version. This makes a specific version of a secret inaccessible, which can be used to revoke access to old credentials immediately. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.enable

Enable a secret version. This allows a previously disabled secret version to be accessed again if necessary for a specific operation. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.get

View secret version metadata. This allows the system to check the state (enabled/disabled) of a secret version before attempting to access it. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.list

List secret versions. This provides a history of the values stored in a secret, allowing for management of version lifecycle and cleanup. This runtime permission is needed at the monitored project level, not at the outpost project.

roles/bigtable.reader

Read all data and metadata from Bigtable tables. This permission enables Data Security Posture Management (DSPM) scanners to access and classify data residing in Bigtable instances within the customer environment.

roles/cloudkms.cryptoKeyEncrypterDecrypter

Encrypt and decrypt data using Cloud KMS keys. This is essential for the DSPM scanner VM to handle encrypted data securely during scanning operations and to ensure communication artifacts remain protected.

roles/cloudkms.viewer

Read metadata of cryptographic keys and key rings. This allows the scanner VM to identify and validate the keys required for encryption operations, ensuring the correct keys are used for data protection.

roles/cloudsql.client

Connect to and execute data operations on Cloud SQL databases. This allows the scanner to authenticate and interact with Cloud SQL instances to perform security assessments and data classification.

roles/secretmanager.secretAccessor

Read secret values from Secret Manager. This permission allows the DSPM scanner or container registry scanner to retrieve credentials needed to access Cloud SQL databases, customer container registries, or other protected resources.

roles/servicenetworking.serviceAgent

Manage private service networking connections. This is required to establish Private Service Access, enabling DSPM to connect securely to Cloud SQL instances.

secretmanager.secrets.create

Create a new secret. This is used to securely store credentials or configuration data required by the scanner, such as temporary database passwords, to connect to scan targets.

secretmanager.secrets.delete

Delete a secret. This allows the system to clean up and remove secrets once they are no longer needed, preventing sensitive data from persisting in the environment.

secretmanager.secrets.get

View secret metadata. This allows the system to verify the configuration and existence of secrets used by the scanner without revealing the payload.

secretmanager.secrets.list

List secrets. This provides visibility into the secrets managed within the project, allowing the system to identify those related to scanning operations.

secretmanager.versions.add

Add a new secret version. This allows the system to update the stored secret value, for example, when rotating temporary credentials for a new scan.

secretmanager.secrets.update

Update secret metadata. This is used to modify labels or settings on secrets, such as marking them for replication or tracking ownership. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.access

Access secret payloads. This permission allows the scanner to retrieve the actual sensitive value (e.g., a password) stored in a secret version to authenticate against a target. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.destroy

Destroy a secret version. This permanently removes a specific version of a secret payload, ensuring that old credentials cannot be recovered. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.disable

Disable a secret version. This makes a specific version of a secret inaccessible, which can be used to revoke access to old credentials immediately. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.enable

Enable a secret version. This allows a previously disabled secret version to be accessed again if necessary for a specific operation. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.get

View secret version metadata. This allows the system to check the state (enabled/disabled) of a secret version before attempting to access it. This runtime permission is needed at the monitored project level, not at the outpost project.

secretmanager.versions.list

List secret versions. This provides a history of the values stored in a secret, allowing for management of version lifecycle and cleanup. This runtime permission is needed at the monitored project level, not at the outpost project.

roles/secretmanager.secretAccessor

Read secret values from Secret Manager. This permission allows the DSPM scanner or container registry scanner to retrieve credentials needed to access Cloud SQL databases, customer container registries, or other protected resources.