MLS-C01 · Question #181
A machine learning (ML) specialist is administering a production Amazon SageMaker endpoint with model monitoring configured. Amazon SageMaker Model Monitor detects violations on the SageMaker endpoint
The correct answer is B. Run the Model Monitor baseline job again on the new training set.. Amazon SageMaker Model Monitor compares live traffic statistics against a pre-computed baseline. When you retrain the model on a new dataset that better represents current production traffic, the statistical properties of the new training data change. However, the old baseline -
Question
A machine learning (ML) specialist is administering a production Amazon SageMaker endpoint with model monitoring configured. Amazon SageMaker Model Monitor detects violations on the SageMaker endpoint, so the ML specialist retrains the model with the latest dataset. This dataset is statistically representative of the current production traffic. The ML specialist notices that even after deploying the new SageMaker model and running the first monitoring job, the SageMaker endpoint still has violations. What should the ML specialist do to resolve the violations?
Options
- AManually trigger the monitoring job to re-evaluate the SageMaker endpoint traffic sample.
- BRun the Model Monitor baseline job again on the new training set.
- CDelete the endpoint and recreate it with the original configuration.
- DRetrain the model again by using a combination of the original training set and the new training
How the community answered
(30 responses)- A3% (1)
- B83% (25)
- C10% (3)
- D3% (1)
Explanation
Amazon SageMaker Model Monitor compares live traffic statistics against a pre-computed baseline. When you retrain the model on a new dataset that better represents current production traffic, the statistical properties of the new training data change. However, the old baseline - created from the original training set - remains in place. Model Monitor will continue to flag violations because it is still comparing against the outdated statistical distribution. The fix is to re-run the baseline job on the new training dataset so that the constraints and statistics reflect the new data. Option A (manually triggering the monitoring job) does not help because the underlying baseline is still stale. Option C (deleting and recreating the endpoint) is unnecessary and destructive. Option D (combining datasets) would not fix the baseline mismatch.
Topics
Community Discussion
No community discussion yet for this question.