How to Configure Vertex AI Model Monitoring for Multiple Models with Traffic Splitting?

You developed a custom model by using Vertex AI to predict your application's user churn rate. You are using Vertex AI Model Monitoring for skew detection. The training data stored in BigQuery contains two sets of features - demographic and behavioral. You later discover that two separate models trained on each set perform better than the original model. You need to configure a new model monitoring pipeline that splits traffic among the two models. You want to use the same prediction-sampling-rate and monitoring-frequency for each model. You also want to minimize management effort. What should you do?

  1. Keep the training dataset as is. Deploy the models to two separate endpoints, and submit two Vertex AI Model Monitoring jobs with appropriately selected feature-thresholds parameters.
  2. Keep the training dataset as is. Deploy both models to the same endpoint and submit a Vertex AI Model Monitoring job with a monitoring-config-from-file parameter that accounts for the model IDs and feature selections. Source Reference Answer
  3. Separate the training dataset into two tables based on demographic and behavioral features. Deploy the models to two separate endpoints, and submit two Vertex AI Model Monitoring jobs.
  4. Separate the training dataset into two tables based on demographic and behavioral features. Deploy both models to the same endpoint, and submit a Vertex AI Model Monitoring job with a monitoring-config-from-file parameter that accounts for the model IDs and training datasets.

Community Votes

B
68%
D
32%

68% of anonymous learners picked answer B. Votes are pick records left by other test-takers — they are not the verified answer.

Community Insight

The exam tests whether you know that Vertex AI Model Monitoring can detect skew per feature from a single training table and that monitoring-config-from-file allows per-model configuration without separating data or creating separate endpoints.

For splitting traffic between two Vertex AI models while minimizing management effort, the community recommends deploying both models to the same endpoint and using monitoring-config-from-file. Keeping the original training dataset intact is preferred because Vertex AI Model Monitoring can analyze features individually for skew detection.

Choosing D, which separates the training dataset into two tables, because it seems logical to give each model its own training data. This is unnecessary for skew detection and adds management overhead, contradicting the requirement to minimize management effort.

Community Discussion (10 comments)

guilhermebutzke 👍 7 Selected: B
My answer: B If you're using Vertex AI Model Monitoring for skew detection and your data is stored in BigQuery, it's not strictly necessary to separate the data into two tables. Vertex AI Model Monitoring can indeed analyze each feature individually to detect skew. So, isn't necessary to separate data. Then, the monitoring-config-from-file parameter lets you specify unique configurations for each model, including ID and training data information. This ensures targeted monitoring and analysis and a unique monitoring job.
Dirtie_Sinkie 👍 1 Selected: D
My vote is D, have to separate the training dataset
bfdf9c8 👍 1 Selected: D
The question mentions skew, yo need to configure the model monitoring with this in mind, so the better option is to separate in two diferent tables to user skew detection
fitri001 👍 3 Selected: B
Reduced Management Effort: You only need to deploy and monitor a single endpoint, minimizing complexity compared to managing two separate endpoints and monitoring jobs (Option A and C). Efficient Data Usage: Maintaining the original training dataset simplifies data management and avoids the need to split it into separate tables (Option C and D). Granular Monitoring: The monitoring-config-from-file parameter allows you to specify configurations for each model within the same monitoring job. You can define the model ID and the features to monitor for potential skew or drift for each model independently.
pinimichele01 👍 1 Selected: B
I don't understand why it is necessary to separate dataset when there is Vertex AI Monitoring
shuvs 👍 1 Selected: D
Not B, as training on separate datasets is recommended.
Yan_X 👍 1 Selected: D
D Separate data to 2 tables to make sure both models are trained with most relevant data.
b1a8fae 👍 1 Selected: D
D. You need to split the training dataset for each respective model. Furthermore, you only need to control for 2 differences between models in monitoring-config-from-file: model ID, and training set. Feature selection should be the same in both models.
shadz10 👍 2 Selected: D
D - makes more sense two models to be trained seperately and more accuarately also submits a Vertex Al Model Monitoring job with a monitoring-config-from parameter which would enable the skew detecttion to work for each model
pikachu007 👍 4 Selected: B
A. Separate Endpoints: This approach involves more management overhead and potentially complicates monitoring configurations. C. Separate Datasets: Splitting the dataset into two tables is unnecessary for model monitoring and could introduce data management complexities. D. Separate Datasets, Same Endpoint: While feasible, this option lacks the flexibility of granular feature control provided by monitoring-config-from-file.

Comments & Corrections

No comments yet — spotted an error or have a note? Share it below.

Log in to comment, report an error, or add a note about this question.

Submitted for moderation before publishing. Keep it helpful and respectful.

Expert Analysis

Why the Answer Is Correct

Option B is correct because it minimizes management effort by deploying both models to a single endpoint and submitting one Vertex AI Model Monitoring job. The monitoring-config-from-file parameter lets you specify unique configurations for each model, including model IDs and feature selections, which is exactly what you need when traffic is split between two models. As a top-voted comment notes, it is not necessary to separate the training data into two tables because Vertex AI Model Monitoring can analyze each feature individually to detect skew. Another comment highlights that this approach reduces management overhead while still providing granular per-feature monitoring.

Why the Other Options Are Wrong

Options A and C require two separate endpoints and two monitoring jobs, which increases management effort even though the same sampling rate and frequency are used. Option C also unnecessarily splits the training dataset into two tables. Option D uses a single endpoint but still separates the training dataset, which is unnecessary and contradicts the goal of minimal management effort. Furthermore, Option D’s monitoring-config-from-file is described as accounting for training datasets, while the correct approach—as in Option B—should account for model IDs and feature selections to leverage Vertex AI’s per-feature skew detection.

Community Comment Notes

Several commenters initially chose D, arguing that separate training datasets are recommended for models trained on different feature sets. However, the majority recognized that Vertex AI Model Monitoring can operate on the original training dataset without splitting it. One commenter with many upvotes stated, 'Vertex AI Model Monitoring can indeed analyze each feature individually to detect skew,' directly refuting the need to separate data. Another comment emphasized reduced management effort as the key advantage of using a single endpoint and monitoring-config-from-file. The consensus is clear: don't split tables unless the monitoring requirement forces you to.

Official Reference

Exam Strategy

When a question emphasizes minimizing management effort, always favor a single endpoint and a single monitoring job over separate resources. Also remember that Vertex AI Model Monitoring can detect skew per feature from an unsplit training table, so avoid selecting options that split the dataset unless there is an explicit technical need.

Related Analysis

Practice All PMLE Questions

Access 65 questions with complete answers and detailed explanations.

View Full PMLE Practice Test →

← Back to PMLE Study Guide