Question 220

A contractor should work with only the marketing analytics project, not every asset in the company domain. What membership scope fits?

  1. SageMaker Unified Studio project membership
  2. SageMaker domain unit hierarchy
  3. SageMaker Unified Studio domain
  4. Service-linked IAM role
  5. IAM managed versus inline policy
  6. IAM identity policy attachment

Explanation

Correct Answer

A:

Project membership controls collaboration and access within the project's governed workspace; in this scenario, add the contractor to the specific project with authorized membership and check the resulting data permissions.

Incorrect Answers

B:

Domain units group related teams and help structure delegated governance within a domain; this helps with separating business-unit stewardship responsibilities, not limiting an analyst's membership to one approved project.

C:

A Unified Studio domain defines an overarching environment for governed data and analytics workspaces; although useful for establishing a common organizational boundary for governed projects, it leaves the need for limiting an analyst's membership to one approved project unmet.

D:

Service-linked roles are IAM roles predefined for supported AWS services to act on an account's behalf; it would target allowing an AWS managed service to perform its own required operations, while the question is about limiting an analyst's membership to one approved project.

E:

Customer-managed IAM policies can be versioned and reused across roles, unlike role-specific inline policies; its benefit is reusing controlled permissions across multiple data pipeline roles; that is different from limiting an analyst's membership to one approved project.

F:

An identity-based IAM policy grants allowed API actions to a user, group or role subject to other policy checks; the current need is for limiting an analyst's membership to one approved project, not authorizing a data-processing role for a narrow AWS API action.

Question 221

An RDS table holds millions of invoices and requests by invoice_number perform full scans. What design change should be considered?

  1. Redshift sort key selection
  2. DynamoDB global secondary index
  3. DynamoDB composite key
  4. RDS secondary B-tree index

Explanation

Correct Answer

D:

A relational secondary index can accelerate selective predicates when its maintenance cost is justified; in this scenario, add an index on the selective lookup column and validate selectivity, write overhead and the optimizer's chosen plan.

Incorrect Answers

A:

Redshift sort keys can improve block elimination for queries filtering on correlated column values; it would target reducing scanned blocks for frequent time-range queries, while the question is about accelerating indexed equality lookups on a relational table.

B:

A GSI supports queries using an alternate partition/sort key without changing the table's primary key; its benefit is querying DynamoDB items by a second business identifier; that is different from accelerating indexed equality lookups on a relational table.

C:

DynamoDB partition and sort keys efficiently organize related items for a known query access pattern; the current need is for accelerating indexed equality lookups on a relational table, not retrieving the latest orders for one account.

Question 222

An order system must update inventory and payment status atomically with foreign-key relationships. Which managed store fits?

  1. Amazon Redshift warehouse
  2. Amazon S3 data lake
  3. Amazon RDS relational database
  4. Amazon DynamoDB table

Explanation

Correct Answer

C:

Amazon RDS manages supported relational engines for transactional workloads using SQL and constraints; in this scenario, a relational engine supports transactional consistency and query patterns for the normalized operational schema.

Incorrect Answers

A:

Redshift offers columnar analytic processing for large SQL aggregations over warehouse datasets; it would target executing repeated analytical joins over large fact tables, while the question is about supporting ACID transactions across related business records.

B:

S3 stores durable objects of varying formats economically without requiring an operational database schema; its benefit is retaining raw petabyte-scale historical source files; that is different from supporting ACID transactions across related business records.

D:

DynamoDB provides managed key-value and document access at low latency using partition keys and indexes; this can support serving predictable millisecond item lookups at high request volume, but would not address the need for supporting ACID transactions across related business records.

Question 223

An S3 notification should trigger validation of a tiny JSON object and append a metadata field. Which compute option is simplest?

  1. Redshift SQL transformations
  2. AWS Glue Spark ETL
  3. Amazon EMR customized frameworks
  4. Lambda per-record enrichment

Explanation

Correct Answer

D:

Lambda is suited to short event-driven transformations within its runtime and resource limits; in this scenario, invoke a Lambda function from the event path for bounded processing and route failures to an appropriate retry mechanism.

Incorrect Answers

A:

Redshift SQL operations can transform warehouse-resident data without exporting it to a separate engine; the current need is for enriching individual small records after an event, not reshaping existing warehouse tables in place.

B:

AWS Glue provides managed Spark-based ETL jobs and integration with the Data Catalog; this can support running recurring serverless-style Spark transformations, but would not address the need for enriching individual small records after an event.

C:

Amazon EMR provides controlled Hadoop and Spark clusters with configurable frameworks and libraries; this helps with running specialized large-scale distributed transformations, not enriching individual small records after an event.

Question 224

A dashboard needs summarized results from Parquet files in S3 without continuously loading a warehouse. Which query layer fits?

  1. Athena result visualization downstream
  2. DataBrew visual profiling
  3. QuickSight SPICE dataset
  4. Amazon QuickSight dashboard

Explanation

Correct Answer

A:

Athena produces query result datasets that can be connected to separate reporting and visualization tools; in this scenario, query cataloged lake data with Athena and connect a supported visualization service to the resulting dataset.

Incorrect Answers

B:

DataBrew's visual interface can inspect column distributions and transformations before publishing cleaned output; this can support exploring raw field problems interactively, but would not address the need for preparing lake queries for a dashboard dataset.

C:

SPICE provides in-memory accelerated storage for supported QuickSight analytic datasets; this helps with speeding repeated dashboard interactions on imported results, not preparing lake queries for a dashboard dataset.

D:

QuickSight builds governed analytic dashboards and visualizations from configured data sources; although useful for publishing metrics to business users through dashboards, it leaves the need for preparing lake queries for a dashboard dataset unmet.

Question 225

A batch orchestration stops after three parallel branches and a Catch handler records an error. What evidence should be inspected?

  1. MWAA worker task log
  2. Step Functions failed state details
  3. Glue job run metrics
  4. MWAA scheduler log

Explanation

Correct Answer

B:

Execution details identify state-level errors and task transitions in a Step Functions workflow; in this scenario, examine the execution path, failed state and error payload to isolate the integration rather than guessing from final status.

Incorrect Answers

A:

Worker task logs record errors raised while an Airflow operator executes; although useful for diagnosing a failed task with running schedulers, it leaves the need for locating which branch and integration caused a state-machine failure unmet.

C:

Glue job run data includes execution duration, failures and relevant service metrics for job diagnosis; its benefit is determining whether an ETL problem is resource-related; that is different from locating which branch and integration caused a state-machine failure.

D:

Scheduler logs show why Apache Airflow DAG runs may not be getting scheduled; the current need is for locating which branch and integration caused a state-machine failure, not diagnosing DAG runs that never enter the queue.

Question 226

Unit tests pass but a deployment changes the Glue job role and breaks its database connection. What additional testing is needed?

  1. CI pipeline integration tests
  2. Immutable deployment artifact
  3. Unit tests with fixture data
  4. Git pull request review

Explanation

Correct Answer

A:

Integration tests exercise real interfaces between components in an isolated environment; in this scenario, run a staging pipeline against test dependencies and verify network access, IAM and end-to-end data contracts.

Incorrect Answers

B:

Versioned packages and immutable images let a pipeline redeploy the exact artifact tested previously; the current need is for validating source-to-target contracts before release, not preventing drift between staging and production ETL code.

C:

Unit tests validate transformation functions against representative inputs and expected results; this can support catching mapping regressions before deployment, but would not address the need for validating source-to-target contracts before release.

D:

Git branches and pull requests support code review before changes merge into a shared baseline; this helps with reviewing ETL changes before production release, not validating source-to-target contracts before release.

Question 227

A Kinesis-triggered Lambda has backlogs although function duration is short; shard-level concurrency is insufficient. Which event-source setting helps?

  1. Lambda batch failure reporting
  2. Kinesis iterator age metric
  3. Kinesis event source mapping
  4. Lambda parallelization factor

Explanation

Correct Answer

D:

The parallelization factor raises concurrent Lambda batches processed per Kinesis shard while preserving partition-key ordering; in this scenario, raise the parallelization factor cautiously and monitor IteratorAge; ordering for each partition key remains preserved.

Incorrect Answers

A:

Lambda can report individual failed record IDs for supported stream batch processing; this can support retrying only failed records in a stream batch, but would not address the need for increasing consumer throughput on busy shards.

B:

IteratorAge measures how far behind a stream-reading Lambda consumer is in processing records; this helps with diagnosing lag in a Lambda stream consumer, not increasing consumer throughput on busy shards.

C:

A Lambda event source mapping reads Kinesis records, checkpoints progress and invokes the function in batches; although useful for connecting a Lambda processor to a Kinesis shard, it leaves the need for increasing consumer throughput on busy shards unmet.

Question 228

A Step Functions-invoked Lambda writes to DynamoDB and currently stores an IAM user key in environment variables. What is preferable?

  1. KMS encryption context policy
  2. IAM role credentials instead of access keys
  3. SSM Parameter Store SecureString
  4. Secrets Manager secret

Explanation

Correct Answer

B:

IAM role sessions provide automatically rotated temporary credentials to supported AWS workloads; in this scenario, use the function's execution role for supported SDK access so no static AWS access key is stored or manually rotated.

Incorrect Answers

A:

KMS key policies can evaluate encryption-context conditions to narrow permitted cryptographic use; its benefit is scoping decrypt permission to the intended encrypted dataset; that is different from authenticating an AWS SDK call without storing an AWS secret.

C:

Systems Manager Parameter Store can hold encrypted configuration values as SecureString parameters; this can support storing a rarely changed encrypted ETL token, but would not address the need for authenticating an AWS SDK call without storing an AWS secret.

D:

Secrets Manager stores and retrieves sensitive credentials and supports managed rotation patterns; this helps with protecting a rotating production database password, not authenticating an AWS SDK call without storing an AWS secret.

Question 229

A Spark-independent data microservice already runs on EKS; its pods saturate CPU at peak. Which Kubernetes control can raise replicas?

  1. Container image multi-stage build
  2. EKS Horizontal Pod Autoscaler
  3. ECS Service Auto Scaling
  4. ECS task CPU and memory sizing

Explanation

Correct Answer

B:

Kubernetes HPA adjusts replica counts of a target workload using metrics and configured policies; in this scenario, configure an HPA with appropriate metrics and resource requests so the Deployment receives additional pods as utilization grows.

Incorrect Answers

A:

Multi-stage Docker builds separate build tools from the runtime image to reduce final image contents; this helps with keeping transformation images lean and faster to distribute, not scaling Kubernetes transformation pods based on utilization.

C:

ECS Service Auto Scaling changes the desired count of running service tasks based on configured metrics; it would target adding parallel workers when a queue backlog rises, while the question is about scaling Kubernetes transformation pods based on utilization.

D:

An ECS task definition reserves CPU and memory for its container workloads; its benefit is giving each transformation container enough compute and RAM; that is different from scaling Kubernetes transformation pods based on utilization.

Question 230

An LLM returns variable JSON keys and occasionally prose, causing downstream schema errors. Which processing control is needed?

  1. Human review for low-confidence extraction
  2. Bedrock text classification request
  3. PII redaction before LLM call
  4. Schema validation of LLM output

Explanation

Correct Answer

D:

Model responses require parsing and validation against an expected structured contract before ingestion; in this scenario, validate fields, types and allowed values; reject or retry malformed responses and retain the raw inference for traceability.

Incorrect Answers

A:

A review queue can route uncertain automated extractions to humans instead of blindly publishing them; the current need is for preventing unstructured model responses from breaking ETL, not avoiding unchecked facts in ambiguous document extraction.

B:

Amazon Bedrock can invoke foundation models to classify or summarize unstructured text in a pipeline; this can support deriving categories from free-form support notes, but would not address the need for preventing unstructured model responses from breaking ETL.

C:

Sensitive fields should be removed or minimized before they are sent to a model where policy requires it; this helps with keeping customer identifiers out of model prompts, not preventing unstructured model responses from breaking ETL.

AWS Certified Data Engineer - Associate DEA-C01 Practice Tests