mediumMultiple Choice
PDE Practice Question: Your team is using Vertex AI Pipelines to…
Your team is using Vertex AI Pipelines to orchestrate a model retraining workflow. The pipeline includes a data validation step, a training step, and a model evaluation step. You want to ensure that if the evaluation step fails due to low model performance, the pipeline stops and does not deploy the model. Which approach should you use?
⚠ Common exam trap
Test-takers frequently confuse retry logic (Option B) with conditional gating, mistakenly thinking that retrying a failed evaluation step will somehow improve model performance, when in fact retries only handle transient errors, not metric-based failures.
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
✓
Use a Conditional in the pipeline to check evaluation metrics and only run the deployment step if metrics pass thresholds
Vertex AI Pipelines supports conditional execution via the `Condition` component, which allows you to evaluate model performance metrics (e.g., accuracy, RMSE) and gate subsequent steps. By placing the deployment step inside a conditional branch that only executes when evaluation metrics meet predefined thresholds, the pipeline automatically stops and avoids deploying a poor-performing model. This approach aligns with MLOps best practices for automated gating in production pipelines.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Run the evaluation step after deployment and roll back if performance is low
Why it's wrong here
Deploying first means a poor model already serves traffic before evaluation runs, so rollback cannot prevent exposure. Evaluation must gate deployment as an upstream pipeline step whose failure halts execution. Post-deployment evaluation suits canary or shadow testing where live traffic validation is deliberately intended.
- ✗
Configure the evaluation step to retry up to 3 times on failure
Why it's wrong here
Retries re-run the same evaluation on identical inputs, so a genuine low-performance result recurs and eventually succeeds or exhausts attempts, allowing deployment. The pipeline needs a conditional that stops on failure. Retries suit transient infrastructure faults, such as a flaky data read, not deterministic model-quality failures.
- ✓
Use a Conditional in the pipeline to check evaluation metrics and only run the deployment step if metrics pass thresholds
Why this is correct
A conditional branch evaluates the evaluation step's output metrics against your thresholds, then gates the deployment step so it only executes when performance passes. This directly satisfies the stem's requirement that a low-performance evaluation halts the pipeline before deployment, since the deployment task is skipped entirely rather than merely flagged.
- ✗
Create a separate pipeline for deployment and trigger it manually after review
Why it's wrong here
A separate manually triggered deployment pipeline decouples evaluation from deployment, so nothing automatically blocks a low-performing model; the gate depends on human diligence. The requirement is automated conditional halting within one pipeline. Separate pipelines suit workflows where deployment timing is deliberately human-controlled.
Go deeper
Related to this question
About these practice questions
One of 747 original PDE practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This PDE 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 PDE exam.