Answer: Monitor the training/serving skew of feature values for requests sent to the endpoint., Enable Vertex Explainable AI feature attribution to analyze model predictions and understand the impact of each feature on the model's predictions.

The question describes a scenario where model performance (AUC) degrades significantly in production after several weeks, indicating potential data drift or concept drift. Option A is correct because monitoring training/serving skew directly addresses data drift, where the statistical properties of production data diverge from the training data, causing performance drops. This is supported by community consensus (67% chose A) and detailed reasoning that Vertex AI Model Monitoring is designed for this purpose. Option C is also correct because enabling Vertex Explainable AI helps diagnose feature-level issues by analyzing feature attributions, revealing if certain features have become less predictive or if their relationships with the target variable have changed over time. Options B and D are less suitable: B (resource utilization) and D (latency) focus on infrastructure performance but do not directly diagnose model performance degradation due to data drift, which is the core issue here.

Author: LeetQuiz Editorial Team