What must you enable on an endpoint for Model Monitor to work?
Why D is the answer
Data capture
Monitoring analyzes captured requests and responses.
In detail
- Data capture records requests and responses for Model Monitor.
- AWS closed SageMaker Model Monitor to new customers. Existing customers can keep using it.
- Captured data goes to S3.
Use case
An illustrative scenario, not a real incident.
A monitoring schedule finds nothing to analyze, because data capture was never enabled.
See it in different ways
sequenceDiagram participant E as Endpoint participant S as Captured data in S3 participant M as Model Monitor participant C as CloudWatch E->>S: Capture requests and responses M->>S: Compare with the baseline M->>C: Publish violations C-->>E: Alarm starts retraining
sequenceDiagram participant E as Endpoint participant S as S3 participant M as Model Monitor participant C as CloudWatch E->>S: Data capture M->>S: Compares with the baseline M->>C: Publishes metrics C-->>M: Alarm on drift
flowchart LR S0["Enable data capture"] --> S1["Requests stored"] --> S2["Monitor analyzes"]
Key point: Monitoring needs data capture.
Why the other options are wrong
- A. Spot training
- Spot is about training cost.
- B. Early stopping
- Early stopping is about training.
- C. Automatic model tuning
- Tuning is about training.