Skip to main content

Use Case for Kendra Index Activity Monitor

  1. Configure Kendra Index Activity Monitor
  2. Use Case for Kendra Index Activity Monitor
  3. Triage Guides by Violation Type
    1. Triage Kendra Index Activity Monitor UNUSED_KENDRA_INDEX

This policy ensures efficient use of AWS Kendra indexing services by monitoring activity and identifying inactive indices. An inactive index incurs unnecessary costs without providing value. By examining the average daily 'IndexQueryCount' metric over a user-defined number of days (default 30), the policy detects indices with no or low usage.

Variables:

  • period: (default 30) The number of days over which to calculate the average requests.
  • requestThreshold: (default 0) The threshold for average requests per day that indicates inactivity.

Pros:

  • Identifies cost-saving opportunities through decommissioning or archiving.

Cons:

  • False positives possible if the index has irregular usage patterns.

How the Policy Works:

  • Fetches the 'IndexQueryCount' metric from CloudWatch for the specified index.
  • Calculates the average requests per day.
  • Compares the average to the threshold.
  • Flags indices with usage below the threshold for possible decommissioning or archiving.

Distinction Between violationTypeIds:

  • UNUSED_KENDRA_INDEX: Triggered when the index's average requests per day fall below the threshold.