Hello everyone,
I’m investigating a CloudTrail logging behavior in QuickSight and would appreciate any insights from the community.
Background
I’m working on implementing a security audit capability to track and monitor data export activities in QuickSight. During my investigation, I’ve identified the following CloudTrail logging behavior:
Current Understanding:
-
File-based data source exports: No API calls are made → No CloudTrail logs generated
-
Database data source exports: API calls are made → CloudTrail logs generated (visible in CloudTrail with eventName: “QueryDatabase” and queryIntent: “EXPORT_TO_CSV” or “EXPORT_TO_EXCEL”)
Unexpected Behavior: When exporting a visualization (dashboard/analysis) that combines both file-based and database data sources, I expected API calls to be made for the database data source portion, which should generate CloudTrail logs. However, no CloudTrail logs are generated at all in this scenario.
Questions
-
What is the underlying mechanism that prevents CloudTrail log generation when exporting visualizations that combine both file-based and database data sources? Why doesn’t the database portion trigger API calls as it would in a standalone database export?
-
Can anyone explain the behavior for both scenarios:
-
Direct Query datasets: When file-based data sources are combined with database data sources
-
SPICE datasets: When file-based data sources are combined with database data sources
Understanding both cases would help establish complete guidelines for security audit capabilities.
-
-
Is there any workaround available to accurately track and monitor data export activities for information security purposes in this situation?
Any guidance or insights would be greatly appreciated. Thank you!