Courseiva

CCNA ML Solution Monitoring, Maintenance, and Security Questions

19 of 94 questions · Page 2/2 · ML Solution Monitoring, Maintenance, and Security · Answers revealed

Practice these questionsDomain overviewAll questions
76
MCQeasy

An organization wants to schedule a retraining pipeline to run every Sunday night. Which AWS service should they use to trigger the pipeline on a schedule?

A.AWS Lambda
B.AWS Step Functions
C.Amazon SQS
D.Amazon EventBridge
AnswerD

Amazon EventBridge provides cron and rate expressions that trigger targets on a defined schedule, satisfying the Sunday-night recurrence requirement without a running server. Its rule-based event bus invokes the pipeline directly, so no polling or manual invocation is needed, and it integrates natively with AWS pipeline services.

Why this answer

Amazon EventBridge is the correct choice because it provides a scheduled event source using cron or rate expressions to trigger target services at specified times, such as every Sunday night. It can directly invoke an AWS Step Functions state machine or a Lambda function to start the retraining pipeline, making it the native scheduling service for event-driven workflows in AWS.

Exam trap

The trap here is that candidates often confuse AWS Lambda's ability to be triggered by a schedule with Lambda itself being a scheduling service, but Lambda is only the compute target, not the scheduler — EventBridge is the service that provides the scheduled trigger.

How to eliminate wrong answers

Option A is wrong because AWS Lambda is a compute service that runs code in response to triggers, but it does not natively provide scheduling capabilities; while you can use Lambda with EventBridge, Lambda alone cannot generate scheduled events. Option B is wrong because AWS Step Functions is a workflow orchestration service that coordinates multiple AWS services, but it does not have built-in scheduling; it requires an external trigger like EventBridge to start execution on a schedule. Option C is wrong because Amazon SQS is a message queue service for decoupling application components, not a scheduling service; it cannot trigger pipelines on a schedule and relies on consumers to poll or receive messages.

Practice this question →
77
MCQmedium

A team is monitoring a SageMaker endpoint and notices that the average latency (ModelLatency) is increasing over time, but the number of invocations is steady. They suspect that the model's inference code is becoming slower due to memory leaks. Which metric should they also examine to confirm this hypothesis?

A.Invocations metric
B.OverheadLatency metric
C.5XXError metric
D.MemoryUtilization metric (if custom)
AnswerD

Steady invocations with rising ModelLatency points to resource exhaustion inside the container rather than traffic growth. MemoryUtilization, where a custom metric is emitted, reveals whether memory is climbing and approaching the instance limit, confirming a leak in the inference code.

Why this answer

MemoryUtilization is the correct metric to examine because memory leaks in the inference code cause gradual memory consumption, which can lead to increased latency due to swapping or garbage collection overhead. Since the number of invocations is steady, the rising latency is likely due to internal resource exhaustion rather than increased load. Monitoring MemoryUtilization (if custom) directly confirms whether memory is growing over time, supporting the hypothesis.

Exam trap

MLA-C01 often tests the distinction between ModelLatency and OverheadLatency, and candidates may incorrectly choose OverheadLatency when the issue is within the model code.

How to eliminate wrong answers

Option A is wrong because Invocations only counts the number of requests, which is already steady and does not indicate memory issues. Option B is wrong because OverheadLatency measures time spent in SageMaker overhead (e.g., request parsing, I/O) not the model's inference code, so it wouldn't confirm a memory leak in the model. Option C is wrong because 5XXError indicates server errors, which would likely appear only after severe memory exhaustion, not as a gradual latency increase.

Practice this question →
78
MCQeasy

An ML engineer wants to be notified when the average inference latency of a SageMaker endpoint exceeds 500 ms for 2 consecutive evaluation periods. Which AWS service combination should they use?

A.CloudWatch Alarm + SNS
B.SageMaker Model Monitor + SNS
C.EventBridge + Lambda
D.SageMaker Clarify + SNS
AnswerA

CloudWatch publishes the endpoint's ModelLatency and OverheadLatency metrics; an alarm with two consecutive evaluation periods and a 500 ms threshold detects the breach, then triggers SNS to notify subscribers. This satisfies both the latency threshold and consecutive-period constraint.

Why this answer

CloudWatch Alarms evaluate metrics over specified periods and can trigger SNS notifications when a threshold is breached for a defined number of consecutive periods. SageMaker endpoints automatically publish inference latency metrics to CloudWatch, so an alarm on the latency metric with a 500 ms threshold and 2 evaluation periods, wired to an SNS topic, delivers exactly the required notification.

Exam trap

MLA-C01 often tests the confusion between Model Monitor (data/model quality) and CloudWatch (infrastructure metrics) — candidates who see 'SageMaker' and 'monitoring' pick Model Monitor, missing that latency is a CloudWatch metric.

How to eliminate wrong answers

Option B is wrong because SageMaker Model Monitor detects data drift, bias, and quality issues in model inputs/outputs — it does not monitor infrastructure latency metrics or send threshold-based alarms. Option C is wrong because EventBridge + Lambda is event-driven automation, not metric-threshold alerting; it lacks the native consecutive-period evaluation that CloudWatch Alarms provide. Option D is wrong because SageMaker Clarify is for bias detection and explainability, not latency monitoring.

Practice this question →
79
Multi-Selectmedium

A company uses SageMaker Model Monitor for feature attribution drift monitoring with SHAP. Which THREE prerequisites must be in place before starting the monitoring schedule? (Select THREE)

Select 3 answers
A.A SageMaker Clarify processing job that computes SHAP values on the captured data
B.Ground truth labels for the inference data
C.Baseline constraints file for data quality
D.Real-time endpoint with data capture enabled
E.A baseline SHAP explainability file from training data
AnswersA, D, E

Clarify runs the SHAP analysis on the current data to compare with baseline.

Why this answer

SageMaker Model Monitor requires a Clarify processing job to compute SHAP values on the captured data as part of the feature attribution drift monitoring setup. This job generates the necessary SHAP explainability values that are compared against the baseline to detect drift.

Exam trap

The trap here is that candidates often confuse the prerequisites for feature attribution drift monitoring with those for data quality monitoring, mistakenly selecting the baseline constraints file (Option C) instead of the baseline SHAP explainability file (Option E).

Practice this question →
80
MCQmedium

A data science team uses Amazon SageMaker Model Monitor to detect data drift in production. They notice that the schema of incoming data (number of features) has changed compared to the training baseline. Which type of monitor is BEST suited to detect this issue?

A.Bias drift monitor
B.Feature attribution drift monitor
C.Data quality monitor
D.Model quality monitor
AnswerC

Data quality monitoring compares incoming data statistics and schema against the training baseline, so a changed feature count is flagged as a schema violation. Model quality and bias monitors assess predictions, not input structure, making them unsuitable here.

Why this answer

The Data quality monitor in SageMaker Model Monitor is specifically designed to detect violations in the input data schema, such as changes in the number of features, feature types, or missing values, by comparing incoming data against a baseline computed from the training dataset. Since the issue is a structural change in the schema (number of features), the Data quality monitor is the correct choice.

Exam trap

The trap here is that candidates confuse 'data drift' (distribution shift) with 'schema change' and incorrectly choose Feature attribution drift monitor, thinking it covers all input changes, but it only tracks importance shifts, not structural feature count violations.

How to eliminate wrong answers

Option A is wrong because Bias drift monitor focuses on detecting bias in model predictions (e.g., demographic parity) and does not monitor input schema changes. Option B is wrong because Feature attribution drift monitor (SHAP-based) tracks changes in feature importance over time, not the number or presence of features. Option D is wrong because Model quality monitor evaluates degradation in prediction accuracy (e.g., AUC, F1) against ground truth, not input data structure.

Practice this question →
81
Multi-Selectmedium

A company is deploying a SageMaker real-time endpoint and needs to monitor inference latency. Which THREE metrics are available from SageMaker for this purpose? (Choose THREE.)

Select 3 answers
A.OverheadLatency
B.Invocations
C.ModelLatency
D.MemoryUtilization
E.Latency
AnswersA, C, E

OverheadLatency measures the time SageMaker spends on request handling outside model inference, such as preprocessing and network overhead. It directly satisfies the stem's inference-latency monitoring requirement, since total latency comprises model latency plus overhead, letting teams isolate endpoint-side delays from model execution time.

Why this answer

Option A, OverheadLatency, is correct because SageMaker publishes it as the time spent on overhead outside the model itself (for example, request routing and response processing) in the endpoint's invocation path, so it directly contributes to observed inference latency. Option C, ModelLatency, is correct because it measures the interval the container spends processing the request, which is the core inference latency component reported by SageMaker. Option E, Latency, is correct because SageMaker reports the total end-to-end time from when the request is received to when the response is returned, which is the primary inference latency metric for a real-time endpoint.

Option B, Invocations, is not a latency metric; it counts the number of requests sent to the endpoint. Option D, MemoryUtilization, is a resource-utilization metric for the instance/container, not an inference latency measurement.

Exam trap

The trap here is that candidates often confuse Invocations (a request count metric) or MemoryUtilization (a resource utilization metric) with latency metrics, but SageMaker specifically provides three distinct latency-focused metrics: Latency, ModelLatency, and OverheadLatency.

Practice this question →
82
MCQmedium

A machine learning engineer is deploying a model to a SageMaker real-time endpoint that must be accessible only from within a specific Amazon VPC and must not have a public IP address. The engineer also needs to ensure that all data in transit between the endpoint and the calling application is encrypted. Which configuration should the engineer use?

A.Deploy the model to a SageMaker endpoint and attach an IAM resource policy that allows only principals from the VPC to invoke it.
B.Deploy the model to a SageMaker endpoint configured with a VPC configuration specifying private subnets and a security group, and invoke it through an interface VPC endpoint.
C.Deploy the model to a SageMaker endpoint with network isolation enabled and use an interface VPC endpoint (AWS PrivateLink) for invocation.
D.Deploy the endpoint with a public IP and use an AWS WAF web ACL to restrict access to the VPC CIDR range.
AnswerB

Configuring the endpoint with a VPC configuration places the endpoint's elastic network interfaces in the specified private subnets, giving it private IP addresses and no public exposure. Invoking through an interface VPC endpoint (AWS PrivateLink) keeps traffic within the AWS network and supports TLS encryption in transit.

Why this answer

To make a SageMaker endpoint private and accessible only within a VPC, the endpoint must be deployed with a VPC configuration that specifies private subnets and security groups, which places its network interfaces in the VPC. Invoking it through an interface VPC endpoint (AWS PrivateLink) ensures traffic stays private and encrypted in transit.

Exam trap

The trap here is confusing IAM-based access control with network-level isolation, assuming that restricting who can invoke the endpoint also makes it private on the network.

Practice this question →
83
MCQeasy

A machine learning engineer needs to monitor a SageMaker endpoint for data drift and receive alerts when drift is detected. They want to use a fully managed AWS service to schedule the monitoring jobs and send notifications. Which AWS service should they use to orchestrate the monitoring schedule and trigger alerts?

A.SageMaker Model Monitor with a monitoring schedule
B.AWS Lambda
C.Amazon CloudWatch
D.Amazon EventBridge
AnswerA

SageMaker Model Monitor allows you to create a monitoring schedule that automatically runs monitoring jobs at specified intervals. It integrates with Amazon CloudWatch to emit metrics and can trigger alarms or notifications via Amazon SNS when violations are detected. This is the managed service designed for this purpose.

Why this answer

SageMaker Model Monitor provides the ability to create a monitoring schedule that runs on a recurring basis. It evaluates data against a baseline and publishes results to CloudWatch, where you can set alarms to send notifications. The other services can be part of the notification chain, but the core scheduling and drift detection is handled by SageMaker Model Monitor.

Exam trap

The trap here is thinking that a general-purpose service like EventBridge or CloudWatch is responsible for creating and scheduling the monitoring jobs, when actually SageMaker Model Monitor manages the schedule.

Practice this question →
84
MCQmedium

A company uses SageMaker Model Monitor to detect bias drift in their real-time inference endpoint. They have collected ground truth labels and want to monitor for bias across different demographic groups. Which type of monitoring should they configure?

A.SageMaker Model Monitor – Feature Attribution Drift Monitoring
B.SageMaker Clarify – Bias Drift Monitoring
C.SageMaker Model Monitor – Model Quality Monitoring
D.SageMaker Model Monitor – Data Quality Monitoring
AnswerB

SageMaker Clarify bias drift monitoring compares live inference data against ground truth labels, computing bias metrics across demographic groups and alerting when they deviate from baseline. This satisfies the requirement to detect bias drift for specific groups.

Why this answer

SageMaker Clarify – Bias Drift Monitoring is the correct choice because it specifically monitors bias metrics over time by comparing ground truth labels with model predictions across different demographic groups. It detects changes in bias metrics such as disparate impact, which is exactly what the company needs.

Exam trap

MLA-C01 often tests the distinction between different Model Monitor types, and candidates may confuse bias drift with feature attribution drift or model quality monitoring.

How to eliminate wrong answers

Option A is wrong because Feature Attribution Drift Monitoring detects changes in feature importance, not bias across groups. Option C is wrong because Model Quality Monitoring tracks metrics like accuracy, precision, and recall, but does not specifically monitor bias. Option D is wrong because Data Quality Monitoring checks for data drift in input features, not bias in predictions relative to ground truth.

Practice this question →
85
MCQhard

A financial services company is deploying a fraud detection model on SageMaker. To comply with regulations, they must ensure that the model's predictions are not biased against protected groups. They plan to monitor bias drift post-deployment using SageMaker Clarify. Which data inputs are required to configure Clarify's bias drift monitoring?

A.Only the inference data with predictions
B.Only the ground truth labels for recent predictions
C.Only the training data with feature attributions
D.Baseline training data with ground truth labels and inference data with predictions
AnswerD

Clarify's bias drift monitoring compares the baseline distribution against live inference data, so it needs labelled baseline training data (ground truth) plus inference data containing predictions. Without both, it cannot compute bias metrics or detect drift across protected groups.

Why this answer

SageMaker Clarify's bias drift monitoring requires a baseline—specifically, the training data with ground truth labels—to establish the original bias metrics, and the inference data with predictions to compute post-deployment bias metrics. By comparing these two datasets, Clarify detects statistically significant shifts in bias over time, which is essential for regulatory compliance in fraud detection models.

Exam trap

The trap here is that candidates often assume only inference data is needed for monitoring, overlooking the critical requirement of a baseline training dataset with ground truth labels to measure drift against.

How to eliminate wrong answers

Option A is wrong because inference data with predictions alone lacks a baseline for comparison, making it impossible to measure drift from the original model behavior. Option B is wrong because ground truth labels for recent predictions, without a baseline training dataset, cannot establish the initial bias metrics needed for drift detection. Option C is wrong because training data with feature attributions, while useful for explainability, does not include the inference data with predictions required to compute post-deployment bias metrics.

Practice this question →
86
MCQeasy

A company wants to reduce costs for a SageMaker real-time endpoint that receives predictable traffic patterns: high during business hours and low at night. The model is a small PyTorch model. Which cost-saving strategy is most suitable?

A.Use a single large instance to handle peak load
B.Use a multi-model endpoint with multiple models
C.Configure auto-scaling with a scheduled scaling policy to add instances during business hours and reduce at night
D.Switch to batch transform jobs and run nightly
AnswerC

Scheduled scaling aligns endpoint instance counts with the predictable daytime peak and nightly trough, so capacity is added only when needed. This directly addresses the stem's cost-reduction goal for a small PyTorch model on a real-time endpoint.

Why this answer

Scheduled auto-scaling is designed for predictable traffic patterns: you define a schedule to scale out during business hours and scale in at night, matching capacity to demand and minimizing cost. For a small PyTorch model on a real-time endpoint, this is the most cost-effective and operationally simple strategy.

Exam trap

The trap is over-engineering with multi-model endpoints or batch transform when the question explicitly states predictable traffic — scheduled auto-scaling is the textbook answer for predictable patterns.

How to eliminate wrong answers

Option A is wrong because using a single large instance sized for peak load wastes money during off-peak hours and does not scale dynamically. Option B is wrong because multi-model endpoints are for hosting multiple models on one endpoint to share resources, not for handling predictable traffic variations of a single model. Option D is wrong because batch transform is for offline, non-real-time inference and cannot serve real-time endpoint traffic.

Practice this question →
87
MCQhard

A media company stores training data in an S3 bucket encrypted with an AWS KMS customer-managed key. A SageMaker training job runs inside a private VPC subnet with no internet access and must read that bucket. The job currently fails with an access-denied error from S3. Which change most directly resolves the failure while preserving the private-network requirement?

A.Enable SageMaker network isolation on the training job and grant the execution role kms:Decrypt on the customer-managed key.
B.Move the training data into an Amazon EFS file system mounted in the subnet and grant the execution role elasticfilesystem:ClientMount permissions.
C.Attach the AmazonS3FullAccess managed policy to the SageMaker execution role and disable the bucket's default encryption temporarily during training.
D.Add an interface VPC endpoint for S3 in the subnet and add the endpoint's condition to the bucket policy and KMS key policy so the endpoint can access both resources.
AnswerD

With no internet access, traffic to S3 must traverse a VPC endpoint, and both the S3 bucket policy and the KMS key policy must permit the endpoint so that authorization succeeds for the encrypted objects. Using a gateway endpoint for S3 also works in many designs, but the interface endpoint with proper policy conditions is the pattern that satisfies both private connectivity and key-based decryption.

Why this answer

A training container in a subnet without internet access can only reach S3 through a VPC endpoint, and encrypted objects additionally require that the endpoint be authorized in both the S3 bucket policy and the KMS key policy. Supplying that endpoint plus the matching key and bucket policy conditions restores the data path while keeping all traffic inside the VPC and the data encrypted.

Exam trap

The trap here is treating an access-denied error as purely an IAM permissions issue, when the real blocker is the missing private network route to S3 plus KMS key policy authorization.

Practice this question →
88
MCQmedium

A team has deployed a real-time inference endpoint and wants to automatically scale based on CPU utilization. Which scaling policy type should they use with Application Auto Scaling for SageMaker endpoints?

A.Target tracking scaling
B.Step scaling
C.Predictive scaling
D.Simple scaling
AnswerA

Target tracking scaling adjusts capacity automatically to hold a chosen metric, such as CPU utilisation, at a specified target value. For SageMaker real-time endpoints, Application Auto Scaling creates the required CloudWatch alarms and scales instances in or out, directly satisfying the stem's requirement to scale on CPU utilisation without manually defining thresholds.

Why this answer

Target tracking scaling is correct because it adjusts capacity to keep a specified metric, such as CPU utilization, at a target value. Application Auto Scaling for SageMaker endpoints supports target tracking, which automatically creates and manages the necessary CloudWatch alarms and scaling policies. This is the recommended approach for maintaining a utilization target without manually defining step adjustments.

Exam trap

The trap is assuming that any scaling policy type works with SageMaker endpoints, when in fact only target tracking and step scaling are supported, and target tracking is the default recommendation for metric-based scaling.

How to eliminate wrong answers

Option B is wrong because step scaling requires you to define explicit scaling adjustments based on alarm breaches, which is more manual and not ideal for simply maintaining a CPU target. Option C is wrong because predictive scaling is not a supported policy type for SageMaker endpoint auto scaling; it is used for EC2 Auto Scaling to forecast load. Option D is wrong because simple scaling is a legacy policy type that waits for a cooldown before responding and is not the recommended way to track a target metric.

Practice this question →
89
Multi-Selectmedium

A machine learning team needs to automatically retrain a model when concept drift is detected in the deployed endpoint's predictions. Which TWO steps should they take? (Choose TWO.)

Select 2 answers
A.Schedule retraining with Amazon EventBridge on a fixed schedule
B.Create a CloudWatch alarm on a model quality metric (e.g., accuracy) and trigger a Lambda function to start a retraining job
C.Set up SageMaker Model Monitor - Model Quality Monitor to compute prediction quality metrics against ground truth
D.Configure SageMaker Model Monitor - Data Quality Monitor to detect input drift
E.Use SageMaker Clarify to monitor bias drift
AnswersB, C

A CloudWatch alarm on a model quality metric detects concept drift by monitoring live prediction accuracy against ground truth. Alarm state-change events invoke Lambda, which programmatically starts a SageMaker retraining job, satisfying the automatic retraining requirement without manual intervention.

Why this answer

Option B is correct because a CloudWatch alarm on a model quality metric such as accuracy can detect when the deployed model's performance degrades, and its alarm action can invoke a Lambda function that programmatically starts a SageMaker retraining job, giving the automatic, event-driven retraining the team requires. Option C is correct because SageMaker Model Monitor's Model Quality Monitor computes prediction quality metrics (accuracy, precision, recall, etc.) by comparing captured endpoint predictions against ground truth labels, which is exactly the mechanism needed to detect concept drift in the model's predictions. Option A is not appropriate because a fixed EventBridge schedule retrains regardless of whether drift actually occurred, which is time-based rather than drift-triggered.

Option D is incorrect because the Data Quality Monitor detects input/data drift (changes in the feature distribution), not concept drift in prediction quality. Option E is incorrect because SageMaker Clarify is used for bias detection and explainability, not for monitoring concept drift in model predictions.

Exam trap

MLA-C01 often tests the confusion between data drift (input distribution shift, detected by Data Quality Monitor) and concept drift (prediction quality degradation, detected by Model Quality Monitor) — candidates who pick Data Quality Monitor miss the 'predictions' wording in the question.

Practice this question →
90
MCQeasy

A company wants to reduce costs for a production SageMaker endpoint that has predictable traffic patterns. They have purchased a Savings Plan. What additional step can they take to further optimize costs while maintaining performance?

A.Use SageMaker Inference Recommender to right-size the endpoint
B.Reduce the number of instances to one, regardless of load
C.Switch from real-time to batch inference
D.Disable auto-scaling
AnswerA

Inference Recommender profiles the model against candidate instance types and configurations, identifying the cheapest instance that still meets latency and throughput targets. This right-sizes the endpoint, so the Savings Plan discount applies to a smaller, better-matched instance, compounding the cost reduction.

Why this answer

SageMaker Inference Recommender runs load tests against candidate instance types and configurations to identify the most cost-effective endpoint that still meets latency and throughput requirements. Since the Savings Plan already discounts compute, right-sizing the underlying instance fleet is the remaining lever for cost optimization without sacrificing performance. This directly addresses the 'maintaining performance' constraint that rules out simply shrinking capacity.

Exam trap

MLA-C01 often tests the misconception that a Savings Plan alone fully optimizes cost — candidates forget that right-sizing the instance fleet via Inference Recommender is the complementary step that preserves performance.

How to eliminate wrong answers

Option B is wrong because reducing to a single instance regardless of load ignores the predictable traffic patterns and will cause throttling, latency spikes, or 5xx errors during peak periods, violating the performance requirement. Option C is wrong because switching from real-time to batch inference changes the application's serving model entirely — it is not a cost optimization for an existing real-time endpoint and would break any synchronous client integration. Option D is wrong because disabling auto-scaling removes the ability to match capacity to demand, which typically increases cost during idle periods or degrades performance during peaks, the opposite of optimization.

Practice this question →
91
MCQmedium

A data scientist uses SageMaker Model Monitor to track feature attribution drift. Which technique does SageMaker Model Monitor use to compute feature attributions?

A.Permutation Feature Importance
B.SHAP
C.Partial Dependence Plots
D.LIME
AnswerB

SageMaker Model Monitor computes feature attributions using SHAP (SHapley Additive exPlanations), which satisfies the requirement for tracking feature attribution drift. The monitor runs a baseline against captured endpoint data, applying SHAP to quantify each feature's contribution, then compares distributions to detect drift in attribution values.

Why this answer

SageMaker Model Monitor uses SHAP (SHapley Additive exPlanations) to compute feature attributions for model explainability and drift detection. SHAP provides a unified measure of feature importance based on cooperative game theory, ensuring consistent and locally accurate attributions across all features.

Exam trap

AWS often tests the misconception that SageMaker Model Monitor uses LIME for explainability because LIME is a popular model-agnostic method, but the service is specifically designed around SHAP for its theoretical properties and integration with the Amazon SageMaker Clarify framework.

How to eliminate wrong answers

Option A is wrong because Permutation Feature Importance measures the drop in model performance when a feature's values are shuffled, but it does not provide per-instance attributions or support the additive feature attribution framework required by SageMaker Model Monitor. Option C is wrong because Partial Dependence Plots show the marginal effect of a feature on the predicted outcome averaged over the dataset, not per-instance feature attributions needed for drift analysis. Option D is wrong because LIME (Local Interpretable Model-agnostic Explanations) approximates the model locally with a simpler surrogate model, but SageMaker Model Monitor specifically integrates SHAP for its theoretical guarantees of consistency and accuracy, not LIME.

Practice this question →
92
MCQeasy

A company wants to track the lineage of their ML models, including the training dataset, hyperparameters, and training job used to produce each model version. Which AWS service should they use?

A.SageMaker ML Lineage Tracking
B.Amazon DynamoDB
C.AWS Glue Data Catalog
D.Amazon S3 object tagging
AnswerA

SageMaker ML Lineage Tracking automatically records relationships between training datasets, hyperparameters, training jobs and resulting model versions, forming a queryable lineage graph. This directly satisfies the requirement to trace each model version back to its originating artefacts, unlike experiment tracking or model registry alone, which store metrics or versions without capturing those dependency links.

Why this answer

SageMaker ML Lineage Tracking is the correct choice because it is purpose-built to record and query the provenance of ML models, capturing relationships between datasets, training jobs, hyperparameters, and model versions. It creates a directed acyclic graph (DAG) of entities (e.g., artifacts, actions, contexts) that allows you to trace how a specific model version was produced, which directly meets the requirement for lineage tracking.

Exam trap

The trap here is that candidates may confuse general-purpose data storage or cataloging services (like DynamoDB or Glue Data Catalog) with the specialized ML lineage tracking service, overlooking that SageMaker ML Lineage Tracking is the only AWS service designed to model the directed relationships between ML artifacts, actions, and contexts.

How to eliminate wrong answers

Option B (Amazon DynamoDB) is wrong because it is a NoSQL key-value and document database designed for low-latency, scalable data storage, not for tracking ML lineage or modeling the complex relationships between training datasets, hyperparameters, and model versions. Option C (AWS Glue Data Catalog) is wrong because it is a metadata repository for data assets (e.g., tables, schemas, partitions) used in ETL and data cataloging, not for capturing the lineage of ML model training runs or hyperparameters. Option D (Amazon S3 object tagging) is wrong because while tags can label S3 objects with metadata like version or dataset name, they cannot capture the relational graph of lineage (e.g., which training job produced which model from which dataset) and lack query capabilities for tracing provenance across multiple artifacts.

Practice this question →
93
MCQhard

A company deploys a model with SageMaker and wants to monitor for concept drift. They have noticed that the relationship between input features and the target variable has changed, causing model accuracy to degrade. However, the input data distribution remains stable. Which type of drift is this, and what is the most appropriate response strategy?

A.Concept drift; ignore the change as long as input distribution remains stable
B.Data drift; update the baseline statistics and continue monitoring
C.Concept drift; retrain the model with newly collected labeled data
D.Data drift; retrain the model with the latest training data
AnswerC

Concept drift occurs when the relationship between input features and the target changes while the input distribution stays stable, exactly as described. Retraining with newly collected labelled data lets the model relearn the updated feature-to-target mapping and restore accuracy.

Why this answer

This is concept drift because the relationship between input features and the target variable has changed while the input data distribution remains stable. The most appropriate response is to retrain the model with newly collected labeled data that reflects the current relationship, as concept drift requires updating the model's learned mapping from features to labels.

Exam trap

The trap here is that candidates confuse concept drift with data drift, assuming any drift requires updating baseline statistics, when in fact concept drift demands retraining with fresh labeled data to realign the model with the new feature-target relationship.

How to eliminate wrong answers

Option A is wrong because ignoring concept drift will cause continued model accuracy degradation, even if the input distribution is stable; concept drift directly impacts predictive performance. Option B is wrong because this is not data drift (input distribution is stable), and updating baseline statistics would not address the changed feature-target relationship. Option D is wrong because data drift refers to changes in input data distribution, not the feature-target relationship, so retraining with the latest training data under the assumption of data drift is a misdiagnosis.

Practice this question →
94
MCQhard

An organization needs to ensure that all data transmitted between containers in a SageMaker training job is encrypted. In the training job configuration, which setting should they enable?

A.Use a KMS key for data encryption
B.Configure the training job in VPC-only mode
C.Enable inter-container traffic encryption
D.Enable network isolation mode
AnswerC

Inter-container traffic encryption secures the peer-to-peer channel between containers within the same training job, using TLS on the network interfaces. Enabling it in the job configuration satisfies the requirement that data transmitted between containers is encrypted, covering traffic that volume-level or at-rest encryption does not protect.

Why this answer

SageMaker training jobs support inter-container traffic encryption, which ensures that data transmitted between containers (e.g., distributed training workers) is encrypted in transit. This setting uses TLS to protect the communication channel, meeting the organization's requirement for encrypted data transmission between containers.

Exam trap

The trap here is that candidates often confuse encryption at rest (KMS keys) with encryption in transit, or assume VPC-only mode or network isolation automatically encrypts inter-container traffic, when in fact they only control network boundaries without enabling TLS encryption between containers.

How to eliminate wrong answers

Option A is wrong because using a KMS key for data encryption applies to data at rest (e.g., EBS volumes or S3 buckets), not to data in transit between containers. Option B is wrong because configuring the training job in VPC-only mode controls network access and routing but does not inherently encrypt inter-container traffic; it only restricts traffic to a VPC. Option D is wrong because enabling network isolation mode prevents the training job from accessing the internet but does not encrypt inter-container communication; it focuses on network segmentation, not encryption.

Practice this question →
← PreviousPage 2 of 2 · 94 questions total

Ready to test yourself?

Try a timed practice session using only ML Solution Monitoring, Maintenance, and Security questions.

Start 20-question session