How many Troubleshooting Scenario Questions questions are on this page?

This page has 8 Troubleshooting Scenario Questions scenario questions for the PMLE exam, each with detailed explanations and wrong-answer analysis.

How should I approach PMLE scenario questions?

Read the full scenario before looking at the answer options. Identify the constraint or requirement in the scenario, then eliminate options that are generally true but wrong for this specific case. Scenario questions reward careful reading over pattern matching.

← Back to Google Professional Machine Learning Engineer questions

Scenario-based practice

Troubleshooting Scenario Questions

Practise Google Professional Machine Learning Engineer practice questions — original exam-style scenarios covering every exam domain, with detailed explanations, wrong-answer analysis, and common exam traps.

Start full practice test Read exam guide

scenario questions

PMLE

exam code

Google Cloud

vendor

Scenario guide

How to approach troubleshooting scenario questions

These questions describe a network symptom and ask you to identify the root cause or the correct fix. They appear across all certification exams and reward systematic thinking over memorisation. The best candidates follow a consistent troubleshooting framework even under time pressure.

Quick answer

Troubleshooting Scenario Questions questions test whether you can apply the concept in context, not just recognise a definition.

How the topic appears in realistic exam-style scenarios.

Which detail in the question changes the correct answer.

How to eliminate plausible but wrong options.

How to connect the question back to the wider exam objective.

Practice scenarios

Question 1mediummultiple choice

Full question →

A data scientist trains an XGBoost model on Vertex AI with a custom container. The model performs well on a held-out test set but fails to generalize in production. They suspect data leakage between training and validation. What is the best practice to prevent this?

A
Store and serve features using Vertex AI Feature Store with point-in-time correctness
Vertex AI Feature Store with point-in-time correctness retrieves feature values as they existed at each training timestamp, preventing future information leaking into training rows. This directly addresses the leakage causing poor generalisation, which a held-out test set alone cannot expose.
B
Implement feature engineering in Vertex AI Pipelines to ensure temporal ordering
Why it fails: Pipelines orchestrate but don't enforce feature consistency by themselves.
C
Store all features in BigQuery and join on timestamp during training and serving
Why it fails: Manual joins are error-prone and may still introduce leakage.
D
Use Vertex AI AutoML instead of custom training
Why it fails: AutoML may not solve custom feature engineering leakage.

Full breakdown with real-world context →

Question 2hardmultiple choice

Full question →

A data science team has trained a TensorFlow model on-premises using a large dataset. When they try to deploy the model to Vertex AI for online predictions, the deployed model fails to start with a ‘MemoryError’. The model artifact is 2 GB, and the machine type is n1-standard-4 (15 GB RAM). What is the most likely cause?

A
The model is stored in a regional bucket and the Vertex AI endpoint is in a different region.
Why it fails: Cross-region bucket and endpoint placement causes latency or access-permission failures, not a MemoryError during model loading. It is tempting because region mismatch is a common Vertex AI deployment pitfall, but that scenario surfaces as storage access errors, not RAM exhaustion at container startup.
B
The machine type does not support TensorFlow models larger than 1 GB.
Why it fails: Vertex AI imposes no fixed 1 GB TensorFlow size ceiling; n1-standard-4 provides 15 GB RAM, so a 2 GB artifact fits. The claim is tempting because machine types do bound deployable model size, but the actual limit derives from available memory, not an arbitrary 1 GB cap.
C
The model is too large for the machine's memory, causing an out-of-memory (OOM) error during loading.
Vertex AI loads the entire 2 GB artifact into the container's RAM before serving. The n1-standard-4 provides 15 GB, but TensorFlow runtime overhead plus deserialisation can exhaust it, so the process is killed with MemoryError rather than a configuration fault.
D
The model file is corrupted or missing dependencies, causing a crash.
Why it fails: A MemoryError at startup indicates the container exhausted its allocated RAM while loading the model, not a corrupt artifact or missing dependency, which would raise import or deserialisation errors instead. Corruption is tempting because it also prevents startup, but the reported symptom is specifically memory exhaustion during model loading.

Full breakdown with real-world context →

Question 3hardmulti select

Full question →

A team is troubleshooting a Vertex AI Pipelines run that keeps failing at the model evaluation step. The pipeline includes steps: data preprocessing, training, evaluation, and deployment. Which THREE actions should they take to diagnose the issue?

A
Verify that the training step output is correctly linked as input to evaluation.
Evaluation fails when its input artifact is missing or malformed. Confirming the training step's output model is correctly wired as the evaluation step's input verifies the pipeline graph passes a valid artifact, ruling out a broken dependency before debugging the code itself.
B
Run the evaluation code locally with the same input data.
Reproducing the evaluation step locally with identical input data isolates whether the failure is code or pipeline infrastructure. If it fails locally, the evaluation logic or data is faulty; if it passes, the issue lies in the Vertex AI Pipelines execution environment.
C
Increase the memory of the evaluation step's machine.
Why it fails: Premature fix without knowing if memory is the issue.
D
Check the logs of the evaluation step in Cloud Logging.
Cloud Logging captures the evaluation step's container stdout and stderr, including the exact exception and stack trace that caused the failure. Reading those logs identifies the precise error, which is faster than guessing at pipeline configuration causes.
E
Replace the evaluation step with a Vertex AI Model Evaluation service.
Why it fails: Changes the process instead of diagnosing the current failure.

Full breakdown with real-world context →

Question 4hardmulti select

Full question →

A company uses Vertex AI Model Monitoring to detect training-serving skew. They have a categorical feature 'product_category' with high cardinality. The monitoring job alerts for skew, but the data scientists believe the model performance is still acceptable. Which THREE actions should the team take to investigate and resolve the alert?

A
Examine which categories have the largest distribution changes to understand the nature of the shift.
Identifying specific categories helps assess whether the drift is due to seasonal effects or other benign causes.
B
Adjust the alerting threshold based on historical drift patterns to reduce noise.
Tuning thresholds helps filter out inconsequential drift.
C
Compare model performance metrics (e.g., AUC) on the drifted segment vs. the non-drifted segment.
Segment-level performance analysis determines if drift is actually harmful.
D
Remove the drifted categories from the feature set to eliminate the alert.
Why it fails: Removing categories reduces model information and could degrade performance.
E
Ignore the alert because the model is performing well; monitoring alerts are often false positives.
Why it fails: Ignoring alerts is not recommended; the team should investigate to confirm it's a false positive.

Full breakdown with real-world context →

Question 5mediummultiple choice

Full question →

A model deployed on Vertex AI Prediction is returning high latency for real-time requests. The model is a small TensorFlow model. Which troubleshooting step should the team take first?

A
Retrain the model with a larger batch size
Why it fails: Batch size affects throughput during training or batch prediction, not per-request latency for a deployed model. Retraining also does not address serving infrastructure. The first step is inspecting the endpoint's machine type and autoscaling. Larger batch sizes are correct for improving training throughput or batch inference efficiency.
B
Check if the machine type is too small and enable autoscaling
A small TensorFlow model should not be inherently slow, so insufficient machine resources or lack of scaling is the likely bottleneck. Checking the machine type and enabling autoscaling addresses CPU or memory saturation before deeper model-level investigation.
C
Use a custom container with optimized runtime
Why it fails: A custom container addresses dependency or framework mismatches, not latency for a small TensorFlow model already served by a prebuilt container. The first step is checking whether the endpoint uses a GPU or CPU machine type and whether traffic is hitting a cold instance. Custom containers suit non-standard runtimes or libraries.
D
Enable Cloud Armor to reduce traffic
Why it fails: Cloud Armor filters and blocks malicious or unwanted HTTP(S) traffic at the edge; it does not reduce legitimate request latency on Vertex AI Prediction. High latency for a small model usually stems from machine type, scaling, or cold starts. Cloud Armor is correct when mitigating DDoS or applying WAF rules to a load balancer.

Full breakdown with real-world context →

Question 6easymultiple choice

Full question →

A machine learning model deployed on Vertex AI is returning erroneous predictions. The team needs to investigate the root cause by examining the prediction request and response details. Which Google Cloud tool is best suited for this?

A
Cloud Monitoring
Why it fails: Cloud Monitoring collects metrics, dashboards and alerting on resource health; it does not retain individual prediction request and response bodies. It is tempting because it is the default observability surface for Vertex AI endpoints, and would be correct for tracking latency, error rates or resource saturation rather than inspecting payload contents.
B
Cloud Debugger
Why it fails: Cloud Debugger inspects live application state via snapshots and logpoints in running code, not the request/response payloads Vertex AI records. It is tempting because it debugs production code without redeploying, which suits diagnosing application logic faults in App Engine or GKE services, but it cannot surface prediction traffic details.
C
Cloud Logging
Cloud Logging captures the raw prediction request and response payloads emitted by Vertex AI endpoints, letting the team inspect exact inputs and outputs to trace erroneous predictions. It satisfies the need to examine request and response details, which aggregate metrics alone cannot expose.
D
Cloud Trace
Why it fails: Cloud Trace records latency and span timing across distributed services, capturing performance data rather than the request and response payloads needed to diagnose wrong predictions. It would be the right tool for profiling slow inference latency, not for inspecting prediction inputs and outputs.

Full breakdown with real-world context →

Question 7hardmultiple choice

Full question →

A data engineer is troubleshooting a Vertex AI Endpoint that serves a large BERT model. After deployment, many prediction requests fail with 'Out of Memory' errors. The machine type is n1-standard-8 (30 GB memory) with no accelerator. Which action will most likely resolve the issue?

A
Change the machine type to n1-highmem-16 (104 GB memory).
The n1-standard-8's 30 GB cannot hold the large BERT model plus inference tensors, causing out-of-memory failures. Moving to n1-highmem-16 raises memory to 104 GB, satisfying the model's footprint without accelerators. More vCPUs alone would not increase memory proportionally.
B
Use batch prediction instead of online prediction.
Why it fails: Batch prediction processes inputs asynchronously from Cloud Storage and returns results to a bucket, so it cannot serve the live endpoint requests failing with out-of-memory errors. It is correct for large offline scoring jobs, not for online serving that requires immediate responses.
C
Add a GPU accelerator (e.g., NVIDIA T4) to offload computation.
Why it fails: GPU helps compute but does not increase available RAM.
D
Quantize the model from FP32 to INT8.
Why it fails: Quantisation shrinks weights and activations, cutting memory use, but it changes numerical precision and requires re-exporting or re-converting the model, which the stem does not mention as available. It is tempting because INT8 is the standard remedy for memory-bound inference, and would be correct if the model could be re-quantised and accuracy loss were acceptable.

Full breakdown with real-world context →

Question 8easymultiple choice

Full question →

Refer to the exhibit. The team notices that the pipeline fails to read data from the specified Cloud Storage path. What is the most likely issue?

Exhibit

pipeline:
  execution_config:
    runner: DataflowRunner
    project: my-project
    region: us-central1
  components:
    - component_type: CsvExampleGen
      component_name: example_gen
      arguments:
        input_basedir: gs://my-bucket/data/

A
The bucket does not exist
Why it fails: A nonexistent bucket would produce a not-found error, yet the stem indicates the path itself is misconfigured rather than the bucket absent. It is tempting because typos in bucket names are common, and would be correct if the error explicitly reported the bucket as missing rather than a path read failure.
B
The pipeline runner is incorrect
Why it fails: The runner executes pipeline code and does not govern access to a Cloud Storage path, so an incorrect runner would not cause a read failure at that location. It is tempting because runner misconfiguration breaks pipelines generally, and would be correct if the pipeline failed to start or could not execute its components at all.
C
The region is mismatched
Why it fails: A region mismatch affects latency and data residency, but Cloud Storage buckets are globally addressable by name, so the pipeline can still read the path. It is tempting because region errors do break some Google Cloud services, and would be correct if the resource were regional, such as a Compute Engine instance or BigQuery dataset.
D
The service account lacks `storage.objectViewer` permission
A missing `storage.objectViewer` role on the service account directly prevents the pipeline from listing and reading objects at the Cloud Storage path, satisfying the stem's read-failure constraint. Without this permission, requests return 403 errors, so the pipeline cannot access the data regardless of path correctness.

Full breakdown with real-world context →

These PMLE practice questions are part of Courseiva's free Google Cloud certification practice question bank. Courseiva provides original exam-style PMLE questions with detailed explanations, topic-based practice, mock exams, readiness tracking, and study analytics.

Troubleshooting Scenario Questions

How to approach troubleshooting scenario questions

Quick answer

Related PMLE topic practice pages

Automating and Orchestrating ML Pipelines practice questions

Serving and Scaling Models practice questions

Architecting Low-Code ML Solutions practice questions

Scaling Prototypes into ML Models practice questions

Collaborating Within and Across Teams to Manage Data and Models practice questions

Monitoring ML Solutions practice questions

Collaborating to manage data and models practice questions

PMLE fundamentals practice questions

PMLE scenario practice questions

PMLE troubleshooting practice questions

Practice scenarios

A data scientist trains an XGBoost model on Vertex AI with a custom container. The model performs well on a held-out test set but fails to generalize in production. They suspect data leakage between training and validation. What is the best practice to prevent this?

A team is troubleshooting a Vertex AI Pipelines run that keeps failing at the model evaluation step. The pipeline includes steps: data preprocessing, training, evaluation, and deployment. Which THREE actions should they take to diagnose the issue?

A model deployed on Vertex AI Prediction is returning high latency for real-time requests. The model is a small TensorFlow model. Which troubleshooting step should the team take first?

A machine learning model deployed on Vertex AI is returning erroneous predictions. The team needs to investigate the root cause by examining the prediction request and response details. Which Google Cloud tool is best suited for this?

A data engineer is troubleshooting a Vertex AI Endpoint that serves a large BERT model. After deployment, many prediction requests fail with 'Out of Memory' errors. The machine type is n1-standard-8 (30 GB memory) with no accelerator. Which action will most likely resolve the issue?

Refer to the exhibit. The team notices that the pipeline fails to read data from the specified Cloud Storage path. What is the most likely issue?

Exhibit