ML Delete Idle Bedrock Models

Opportunity Name:

ML Delete Idle Bedrock Models

AWS Resource Type:

Amazon Bedrock

Opportunity Description:

CloudFix identifies and removes unused Bedrock customized models that have not incurred any inference requests over a specified period. This action eliminates unnecessary storage charges associated with these models, optimizing costs for AWS users.

Criteria for identifying the opportunity:

• Analysis of Bedrock logs to detect models with no inference traffic.
• Models are considered for removal if they have not seen any inference requests for at least 30 days.
• Annual cost, extrapolated from the last 31 days of usage, exceeds the annual public cost threshold (default $100).

Potential savings (range in % on annual basis):

• Customers can expect to save on storage costs for each unused model detected and removed.
• Actual savings will depend on the number of unused models and their associated storage costs.

Can CloudFix apply an automatic fix?

No

Other considerations:

• Performance impact: Removing unused models has no impact on active workloads or inference operations.
• Data loss considerations: Users should ensure that models marked for deletion are genuinely unused. There may be a need to archive model data before deletion if future use is anticipated.
• Security concerns: Ensure that access policies and permissions are reviewed to prevent unauthorized deletion of models.

Additional Resources:

• CloudFix Knowledge Base
• What is Amazon Bedrock?
• Monitoring Amazon Bedrock with Amazon CloudWatch