Google CloudOverview
Feature stores often hold derived customer, behavioral, operational, or business-sensitive data. Online feature data should have retention configured so stale training and inference data does not accumulate indefinitely.
Remediation guidance
Remediation
Configure online feature-store retention based on the model use case and data classification. Align retention with privacy, audit, and retraining needs.
- Classify the feature data.
- Set a retention period for online storage.
- Validate backup, replay, and retraining requirements before deleting historical data.
Query logic
These are the stored checks tied to this control.
AI feature stores should have retention configured
Connectors
Google Cloud
Covered asset types
AI ServicesDataFeature Store
Expected check: eq []
{
vertexAIFeaturestores(where: { onlineStorageTTLDays_LTE: 0 }) { ...AssetFragment }
}
