PMLE Monitoring ML Solutions Practice Question
An MLOps engineer is setting up monitoring for a deployed model on Vertex AI Endpoints. Which TWO actions are required to enable Vertex AI Model Monitoring for feature skew and drift? (Choose two.)
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Enable request/response logging on the Vertex AI Endpoint
To enable model monitoring, you must enable request/response logging on the endpoint (to capture serving data) and create a monitoring job with the desired configuration.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Export ground truth labels to Cloud Storage
Why it's wrong here
Ground truth export supports model quality evaluation, not feature skew or drift, which compare live traffic against a training reference dataset without labels. Exporting labels to Cloud Storage would be correct when configuring continuous evaluation that scores predictions against delayed actual outcomes.
- ✓
Enable request/response logging on the Vertex AI Endpoint
Why this is correct
Model Monitoring computes skew and drift by comparing live serving traffic against the training baseline, so it needs sampled request and response payloads. Enabling request/response logging on the endpoint supplies that input, satisfying the stem's requirement for drift detection.
- ✗
Enable Vertex AI Pipelines to run scheduled monitoring
Why it's wrong here
Vertex AI Pipelines orchestrates ML workflows; it does not provision the monitoring job itself. Enabling monitoring requires configuring the endpoint's monitoring schedule with a reference dataset and skew/drift thresholds. Pipelines would be correct for automating retraining triggered by those alerts.
- ✓
Create a ModelMonitoringJob with a monitoring configuration
Why this is correct
The ModelMonitoringJob is the resource that actually instantiates monitoring against the endpoint, carrying the skew and drift configuration. Without creating it, no monitoring objective is evaluated, so it satisfies the stem's requirement to enable feature skew and drift detection.
- ✗
Deploy the model with an explanation spec
Why it's wrong here
An explanation spec enables feature attributions via Explainable AI, not skew or drift detection. Monitoring needs a training reference dataset plus a monitoring schedule attached to the endpoint. Explanation specs would be correct when the requirement is per-prediction feature attribution for interpretability.
Go deeper
Related to this question
About these practice questions
Courseiva writes every PMLE question from scratch — 775 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This PMLE practice question is part of Courseiva's free Google Cloud certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the PMLE exam.