An IoT company uses Amazon SageMaker to train and test an XGBoost model for object detection. ML engineers need to monitor performance metrics when they train the model with variants in hyperparameter

Question

Options

• AUse Amazon CloudWatch to monitor performance metrics. Use Amazon Simple Queue Service
• BUse Amazon CloudWatch to monitor performance metrics. Use Amazon Simple Notification
• CUse AWS CloudTrail to monitor performance metrics. Use Amazon Simple Queue Service
• DUse AWS CloudTrail to monitor performance metrics. Use Amazon Simple Notification Service

• A
  8% (5)
• B
  85% (51)
• C
  2% (1)
• D
  5% (3)

Explanation

Option B is correct because Amazon CloudWatch natively integrates with SageMaker to capture and visualize training metrics (loss, accuracy, etc.) across hyperparameter variants, and Amazon Simple Notification Service (SNS) is the AWS service designed specifically for push-based messaging including SMS text messages to phone numbers.

Why the distractors fail:

• A (SQS): SQS is a message queue for decoupling services and buffering work - it does not send SMS messages to people; that's SNS's job.
• C (CloudTrail + SQS): CloudTrail records API audit logs (who called what API, when) - it is not a metrics monitoring tool and cannot track ML training performance curves.
• D (CloudTrail + SNS): SNS is correct for SMS, but CloudTrail is still wrong for performance metrics monitoring; only CloudWatch provides metric dashboards and alarms for SageMaker training jobs.

Memory tip: Think "Watch = Metrics, Trail = Audit" - CloudWatch watches your numbers in real time, CloudTrail trails (tracks) user actions for compliance. For notifications, remember SNS = "Send Notifications to Subscribers" (including via SMS), while SQS = "Store/Queue for Services."

No community discussion yet for this question.