This page provides an overview of Sensitive Actions Service, a built-in service of Security Command Center that detects when actions are taken in your Google Cloud organization, folders, and projects that could be damaging to your business if they are taken by a malicious actor.
In most cases, the actions that are detected by Sensitive Actions Service do not represent threats, because they are taken by legitimate users for legitimate purposes. However, the Sensitive Actions Service cannot conclusively determine legitimacy, so you might need to investigate the findings before you can be sure that they don't represent a threat.
How Sensitive Actions Service works
Sensitive Actions Service automatically monitors all of your organization's Admin Activity audit logs for sensitive actions. Admin Activity audit logs are always on, so you do not need to enable or otherwise configure them.
When Sensitive Actions Service detects a sensitive action that is taken by a Google account, Sensitive Actions Service writes a finding to Security Command Center in the Google Cloud console and a log entry to the Google Cloud platform logs.
Sensitive Actions Service findings are classified as observations and can be viewed by finding class or finding source on the Findings tab in the Security Command Center console.
Restrictions
The following sections describe restrictions that apply to Sensitive Actions Service.
Account support
Sensitive Actions Service detection is limited to actions taken by user accounts.
Encryption and data residency restrictions
To detect sensitive actions, Sensitive Actions Service must be able to analyze your organization's Admin Activity audit logs.
If your organization encrypts your logs by using customer-managed encryption keys (CMEK) to encrypt your logs, Sensitive Actions Service cannot read your logs and, consequently, cannot alert you when sensitive actions occur.
Sensitive actions cannot be detected if you have configured the location of the
log bucket for your your Admin Activity Audit Logs to be in a location other
than the global
location. For example, if you have specified a storage
location for the _Required
logs bucket in a certain project, folder, or organization, logs from that
project, folder, or organization cannot be scanned for sensitive actions.
Sensitive Actions Service findings
The following table shows the finding categories that Sensitive Actions Service can produce. The display name for each finding starts with the MITRE ATT&CK tactic that the detected action could be used for.
Display name | API name | Description |
---|---|---|
Defense Evasion: Organization Policy Changed |
change_organization_policy |
An organization-level organization policy was created, updated, or deleted, in an organization that is more than 10 days old. This finding isn't available for project-level activations. |
Defense Evasion: Remove Billing Admin |
remove_billing_admin |
An organization-level billing administrator IAM role was removed, in an organization that is more than 10 days old. |
Impact: GPU Instance Created |
gpu_instance_created |
A GPU instance was created, where the creating principal has not created a GPU instance in the same project recently. |
Impact: Many Instances Created |
many_instances_created |
Many instances were created in a project by the same principal in one day. |
Impact: Many Instances Deleted |
many_instances_deleted |
Many instances were deleted in a project by the same principal in one day. |
Persistence: Add Sensitive Role |
add_sensitive_role |
A sensitive or highly-privileged organization-level IAM role was granted in an organization that is more than 10 days old. This finding isn't available for project-level activations. |
Persistence: Project SSH Key Added |
add_ssh_key |
A project-level SSH key was created in a project, for a project that is more than 10 days old. |
What's next
- Learn about using Sensitive Actions Service.
- Learn how to investigate and develop response plans for threats.