Question 250

Before activating change capture, analysts need all existing rows from a supported transactional database copied to the destination. Which DMS task mode fits?

  1. Amazon AppFlow
  2. S3 object batch processing with Glue
  3. AWS DMS full load
  4. AWS Glue JDBC connection

Explanation

Correct Answer

C:

AWS DMS can copy existing database tables as an initial migration or replication phase; in this scenario, full load initializes target tables from source contents; later CDC can handle subsequent modifications separately.

Incorrect Answers

A:

AppFlow transfers data between supported SaaS applications and AWS destinations with managed connectors; this can support importing CRM objects from a supported SaaS application, but would not address the need for taking a one-time baseline of source tables.

B:

Glue batch ETL processes stored S3 objects when a scheduled job runs against their paths; this helps with reprocessing a monthly collection of stored flat files, not taking a one-time baseline of source tables.

D:

Glue can read JDBC-backed sources through a connection with network and credential settings; it would target extracting scheduled rows from a private relational database, while the question is about taking a one-time baseline of source tables.

Question 251

An integration requires compact sequential records with an explicit writer schema for downstream decoding. Which file format is suitable?

  1. Apache Parquet output
  2. Compression-aware partitioned Parquet
  3. JSON preservation for variable schema
  4. Apache Avro for row-oriented exchange

Explanation

Correct Answer

D:

Avro serializes row-oriented records with schema-based encoding suitable for data interchange; in this scenario, avro carries schema-aware serialized records and is useful for streaming or batch interchange rather than columnar scan optimization.

Incorrect Answers

A:

Parquet stores column-oriented data with encoding and compression suitable for selective analytics; its benefit is reducing scan bytes for repeated analytic queries; that is different from exchanging schema-described records across heterogeneous consumers.

B:

Compression and sensible file sizing improve Parquet read efficiency without creating countless tiny files; the current need is for exchanging schema-described records across heterogeneous consumers, not avoiding a data lake filled with tiny output objects.

C:

JSON can preserve nested and changing attributes that are awkward in a rigid flat-row schema; this can support preserving semi-structured event attributes during initial landing, but would not address the need for exchanging schema-described records across heterogeneous consumers.

Question 252

A private data subnet reaches S3 through a VPC endpoint but should only access approved data-lake buckets. Which additional control applies?

  1. Network ACL rule order
  2. Private VPC endpoint policy
  3. Glue worker security-group egress
  4. Database security-group ingress

Explanation

Correct Answer

B:

VPC endpoint policies can restrict supported API actions and resource access through a particular endpoint; in this scenario, configure an endpoint policy that scopes allowed actions and S3 resources, alongside IAM and bucket policies.

Incorrect Answers

A:

Network ACLs are stateless subnet controls with numbered allow and deny rules evaluated in order; this can support debugging return traffic dropped at a subnet boundary, but would not address the need for limiting S3 API access through a specific gateway endpoint.

C:

Security-group egress rules govern which destinations Glue worker ENIs may contact; although useful for allowing a job to initiate approved outbound database connections, it leaves the need for limiting S3 API access through a specific gateway endpoint unmet.

D:

A VPC security group can restrict database listener access by source security group and destination port; it would target permitting only Glue ETL workers to reach RDS over TCP, while the question is about limiting S3 API access through a specific gateway endpoint.

Question 253

A reconciliation process sees inconsistent totals when concurrent writes occur in its relational source. What database behavior should be reviewed?

  1. Relational isolation-level selection
  2. Redshift lock monitoring
  3. Shorter database transaction
  4. Partitioned load scheduling

Explanation

Correct Answer

A:

Transaction isolation controls the concurrency anomalies a database transaction permits under its engine semantics; in this scenario, evaluate isolation levels and snapshot behavior against workload requirements rather than applying arbitrary retries without consistency design.

Incorrect Answers

B:

Redshift system views and query monitoring expose sessions waiting on resource or transaction locks; its benefit is diagnosing warehouse query waits caused by locks; that is different from choosing consistency guarantees for simultaneous source updates.

C:

Reducing transaction length releases row and table locks sooner in conventional relational engines; the current need is for choosing consistency guarantees for simultaneous source updates, not limiting lock contention from long-running batch updates.

D:

Loading separate table partitions at coordinated intervals can reduce contention with user-facing operations; this can support avoiding bulk-load conflicts with live warehouse queries, but would not address the need for choosing consistency guarantees for simultaneous source updates.

Question 254

A warehouse runs nearly full-time with steady heavy SQL demand and administrators want committed capacity planning. Which model merits evaluation?

  1. EMR persistent cluster
  2. Athena on-demand SQL
  3. Redshift Serverless
  4. Lambda short execution
  5. Provisioned Redshift cluster

Explanation

Correct Answer

E:

Provisioned Redshift compute provides explicit sizing and reservations for more predictable warehouse workloads; in this scenario, size a provisioned cluster or reserved capacity against utilization and compare lifecycle and administration overhead with serverless.

Incorrect Answers

A:

A persistent EMR cluster can run continuing workloads with custom node and framework configurations; it would target maintaining specialized continuously running Spark infrastructure, while the question is about optimizing sustained analytics demand under stable capacity.

B:

Athena can query data in S3 without administering a persistent warehouse cluster; its benefit is serving occasional lake SQL queries with variable demand; that is different from optimizing sustained analytics demand under stable capacity.

C:

Redshift Serverless runs warehouse workloads without manually managing provisioned clusters; the current need is for optimizing sustained analytics demand under stable capacity, not running variable warehouse SQL workloads with less infrastructure management.

D:

Lambda charges for invoked execution and suits bounded tasks under platform runtime constraints; this can support handling sporadic lightweight event transformations without idle servers, but would not address the need for optimizing sustained analytics demand under stable capacity.

Question 255

ETL output occupies twice its expected S3 storage and has repetitive text columns. What file configuration could help?

  1. Parquet compression codec
  2. Relational composite index order
  3. S3 partition cardinality
  4. Redshift distribution key choice

Explanation

Correct Answer

A:

A compression codec changes encoded data size and decompression trade-offs for columnar lake files; in this scenario, compare supported Parquet codecs such as Snappy or ZSTD for the workload's CPU, compatibility and scan cost.

Incorrect Answers

B:

The order of columns in a composite B-tree index affects which query predicates can use its leading keys; the current need is for reducing bytes stored and scanned for Parquet files, not optimizing multi-column predicates in an operational database.

C:

Partitioning by extremely high-cardinality fields can create excessive small partitions and metadata overhead; this can support avoiding millions of tiny lake partitions, but would not address the need for reducing bytes stored and scanned for Parquet files.

D:

Redshift distribution style influences where rows reside and how much data moves during joins; this helps with reducing redistribution for frequent large-table joins, not reducing bytes stored and scanned for Parquet files.

Question 256

A warehouse combines CRM and checkout feeds but the same person appears twice under the shared customer ID. What processing operation is necessary?

  1. Deduplication across sources
  2. Stable join key
  3. Late-arriving dimension handling
  4. Time zone normalization

Explanation

Correct Answer

A:

Entity deduplication uses an agreed business key and conflict-resolution rule across integrated sources; in this scenario, deduplicate on the canonical key with an explicit source-priority or recency rule rather than relying on file order.

Incorrect Answers

B:

Joining data from separate systems requires a trustworthy key that identifies the same entity on each side; this helps with preventing incorrect joins across customer feeds, not merging duplicate customer updates from two systems.

C:

Late-arriving dimensions require a defined strategy for facts received before their lookup rows; although useful for loading facts that precede their dimension records, it leaves the need for merging duplicate customer updates from two systems unmet.

D:

Normalizing timestamps to a common time basis prevents false ordering across systems using different zones; it would target correlating event timestamps from multiple geographic sources, while the question is about merging duplicate customer updates from two systems.

Question 257

A Glue publish job starts before upstream validation completes, exposing partial data. What orchestration control fixes ordering?

  1. EventBridge Scheduler flexible time window
  2. Glue conditional trigger
  3. Airflow dependency trigger rule
  4. Airflow task retry policy
  5. Step Functions execution history

Explanation

Correct Answer

B:

Glue conditional triggers can start downstream jobs when designated upstream jobs or crawlers reach configured states; in this scenario, use a conditional workflow trigger based on successful upstream completion so publish starts only after validation succeeds.

Incorrect Answers

A:

Flexible time windows let a scheduled invocation occur within a configured interval instead of at an exact instant; although useful for spreading noncritical batch launches to avoid thundering herds, it leaves the need for running a publish job only after upstream Glue success unmet.

C:

Airflow trigger rules determine when a downstream task may run relative to upstream task states; its benefit is ensuring a cleanup task runs after upstream failures; that is different from running a publish job only after upstream Glue success.

D:

Airflow tasks support retry counts and delays within a DAG's execution schedule; the current need is for running a publish job only after upstream Glue success, not retrying a transiently failing DAG task.

E:

Step Functions maintains execution event history for supported workflows and helps locate failed states; this can support tracing the precise failure transition in a workflow, but would not address the need for running a publish job only after upstream Glue success.

Question 258

Operators want a page only when a Glue failure metric crosses a threshold over several evaluation periods. Which feature fits?

  1. SNS topic fan-out
  2. CloudWatch alarm action
  3. Step Functions workflow status alert
  4. SQS durable queue
  5. EventBridge event-to-target routing

Explanation

Correct Answer

B:

CloudWatch alarms evaluate metrics against thresholds and can invoke configured notification actions; in this scenario, create a metric alarm with an SNS action and tune evaluation periods to reduce noise from isolated transient events.

Incorrect Answers

A:

SNS publishes notifications to multiple subscribed endpoints or queues using topic subscriptions; although useful for delivering one alert to several independent recipients, it leaves the need for alerting after repeated error metric breaches unmet.

C:

An EventBridge rule can match Step Functions execution status changes for monitoring and notifications; its benefit is notifying when a whole workflow fails rather than one retrying task; that is different from alerting after repeated error metric breaches.

D:

SQS retains messages for asynchronous processing by consumers that may be temporarily unavailable; the current need is for alerting after repeated error metric breaches, not buffering remediation work until a worker can run.

E:

EventBridge rules route matching state-change events to designated consumers; this can support notifying only after a specific job state transition, but would not address the need for alerting after repeated error metric breaches.

Question 259

A Spark join sends nearly all records for one large tenant to one reducer and most workers sit idle. Which skew treatment can help?

  1. Stratified random sampling
  2. Systematic sampling caveat
  3. Spark key salting
  4. Reservoir sampling
  5. Random sampling

Explanation

Correct Answer

C:

Key salting distributes records for a heavily skewed join key across multiple synthetic partitions; in this scenario, add controlled salts and replicate the small-side data as required, then recombine results without changing logical counts.

Incorrect Answers

A:

Stratified sampling draws observations from defined subgroups to preserve rare but important categories; this can support assessing a dataset with very uneven country distributions, but would not address the need for reducing one-key hotspots in a distributed join.

B:

Systematic sampling takes observations at regular intervals and may bias results when the stream is periodic; this helps with avoiding periodic bias when checking machine readings, not reducing one-key hotspots in a distributed join.

D:

Reservoir sampling can maintain a bounded random sample from a stream of unknown total length; it would target sampling a large unknown-length event stream with fixed memory, while the question is about reducing one-key hotspots in a distributed join.

E:

Simple random sampling selects records with equal inclusion probability when an unbiased population estimate is needed; its benefit is building a general quality sample without subgroup constraints; that is different from reducing one-key hotspots in a distributed join.

Question 260

A small Glue job finishes in minutes but consistently reserves many idle workers. What cost optimization should be evaluated?

  1. Column projection before scan
  2. Partition pruning
  3. EMR managed scaling
  4. Glue worker sizing and job profiling

Explanation

Correct Answer

D:

Glue worker types and counts affect job parallelism, memory and billed compute consumption; in this scenario, inspect run metrics, reduce worker allocation where performance allows, and measure duration and DPU consumption together.

Incorrect Answers

A:

Reading only required fields can reduce bytes processed by columnar storage and SQL queries; the current need is for right-sizing an oversized Glue ETL run, not minimizing scanned data during analytic transformations.

B:

Partition pruning excludes dataset partitions whose values cannot satisfy a query predicate; this can support avoiding reads of unrelated months in a data lake, but would not address the need for right-sizing an oversized Glue ETL run.

C:

EMR managed scaling can adjust cluster capacity within configured limits based on workload demand; this helps with reducing idle cluster capacity between Spark stages, not right-sizing an oversized Glue ETL run.

AWS Certified Data Engineer - Associate DEA-C01 Practice Tests