Question

A data science team enables SageMaker Model Monitor with a data quality monitoring schedule on a real-time endpoint. The monitoring jobs run but consistently fail with an error indicating they cannot write violation reports to the designated S3 bucket. The bucket is encrypted with a customer-managed KMS key, and data capture to S3 works correctly. What is the MOST likely cause and appropriate fix?

Accepted Answer

The Model Monitor execution role lacks s3:PutObject and kms:GenerateDataKey permissions for the report bucket and its KMS key; update the role's IAM policy and the KMS key policy to grant access.

Answer

The endpoint's autoscaling policy is throttling the monitoring container; disable target-tracking scaling during monitoring windows.

Answer

The monitoring schedule frequency is set too high; reduce it to hourly so reports have time to write to S3.

Answer

Data capture must be disabled before monitoring reports can be written; turn off endpoint data capture and rerun the schedule.