An auditor asks which service records, for every pipeline run, the input datasets, the hyperparameters, and the resulting model versions, so that any artifact can be traced back to its origins. Which service answers that?
Show answer and explanation
Answer: D
Vertex ML Metadata records executions, artifacts, and contexts, linking each run's inputs, parameters, and outputs. That lineage supports reproducibility and audits. Audit logs document API calls, monitoring tracks metrics over time, and bucket versioning keeps file history, and none of these connects a model back to its exact inputs.
Why the other options are wrong
- A. Audit logs record who called which API and when. They do not describe the datasets, parameters, or artifacts that a training run consumed.
- B. Monitoring shows resource metrics and alerts. Its time series cannot explain which dataset produced a given model version.
- C. Versioning preserves older copies of objects but has no notion of pipeline runs. It cannot link a model to the parameters that created it.