100% Latest & Updated Amazon AWS Certified Data Engineer - Associate DEA-C01 Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!
Question 630
After resolving the following incident, which ongoing safeguard most directly prevents recurrence? A Glue JDBC extraction overloads a small OLTP database when hundreds of partitions open simultaneous connections.
- Automate broker endpoint checks and rolling-client reconnection tests.
- Set source-specific connection budgets and production-safe load tests.
- Require delivery manifests and replay-safe keys for partner batches.
- Keep job-run IDs and manifest IDs linked in operational metadata.
Explanation
Correct Answer
B:Set source-specific connection budgets and production-safe load tests. This safeguards against a recurring cause: Extraction parallelism ignores source capacity and transaction impact.
Incorrect Answers
A:That safeguard addresses another case: An MSK producer succeeds only intermittently after brokers change because its connection string contains one retired broker address. It is not the control for this recurring failure.
C:That safeguard addresses another case: A partner uploads a CSV over SFTP to Transfer Family but a load starts before the file finishes writing. It is not the control for this recurring failure.
D:That safeguard addresses another case: A production Glue workflow receives the same completed-file notification twice and creates two warehouse loads. It is not the control for this recurring failure.
Question 631
A Glue ETL job completes with warnings but the orchestrator assumes success despite a downstream quality check failing. Select TWO complementary measures that correctly repair and validate this failure.
- Convert to validated decimal types upstream or with explicit safe casts before aggregation.
- Only a validated output is promoted to the consumer-facing location.
- Publish all outputs when the ETL runtime exits successfully.
- Add explicit row-count, schema, and quality validation before the publish step.
- A terminated cluster's step and executor logs remain available for diagnosis.
Explanation
Correct Answers
B:Only a validated output is promoted to the consumer-facing location. This is the observable result required before accepting the correction.
D:Add explicit row-count, schema, and quality validation before the publish step. The underlying cause is: Job terminal status is not equivalent to acceptance of the produced data product.
Incorrect Answers
A:That corrective action is for a different case: An Athena SQL query sums decimal prices stored as strings and uses lexicographic ordering on text. It does not resolve this incident.
C:Publish all outputs when the ETL runtime exits successfully. That shortcut conflicts with the stated requirement.
E:This would validate a different case: An EMR job intermittently fails due to capacity interruptions but its logs are stored only on the terminated instance. It does not verify the outcome required here.
Question 632
A Firehose stream delivering transactions to S3 mixes every customer into the same prefix, making Athena queries scan all tenants. Which engineering change best addresses the underlying issue?
- Measure runtime and costs by job size and configure eligible autoscaling or differentiated job profiles.
- Choose a stable order identity and the correct aggregation or distinct-current-state logic.
- Use an appropriate hashed lookup structure while preserving deterministic output behavior.
- Create a supported Firehose dynamic-partitioning delivery stream with well-chosen keys and S3 prefixes.
Explanation
Correct Answer
D:Create a supported Firehose dynamic-partitioning delivery stream with well-chosen keys and S3 prefixes. It directly addresses the case cause (The delivery layout lacks event-derived partition selection.) and should be verified by: A tenant query scans only the relevant partitions and records appear under expected prefixes.
Incorrect Answers
A:That change addresses a Glue job uses a fixed worker count at peak and wastes compute on tiny nightly delta batches; it would not repair this incident.
B:That change addresses a nightly aggregation counts an order once per status-change event rather than once per order ID; it would not repair this incident.
C:That change addresses a large ETL data structure uses a list for repeated membership checks and CPU use grows unexpectedly; it would not repair this incident.
Question 633
A Redshift UNLOAD writes reporting data to an S3 bucket in another account but fails with access denied. Select TWO complementary measures that correctly repair and validate this failure.
- Make the destination bucket world-writable to fix UNLOAD.
- Test similar products rank appropriately for documented benchmark queries.
- A controlled UNLOAD produces output in only the approved S3 prefix.
- Separate or migrate embeddings under explicit model versions and index compatibility rules.
- Grant scoped bucket permissions and necessary KMS access to the Redshift execution role.
Explanation
Correct Answers
C:A controlled UNLOAD produces output in only the approved S3 prefix. This is the observable result required before accepting the correction.
E:Grant scoped bucket permissions and necessary KMS access to the Redshift execution role. The underlying cause is: The UNLOAD role and destination policy do not authorize the intended cross-account write.
Incorrect Answers
A:Make the destination bucket world-writable to fix UNLOAD. That shortcut conflicts with the stated requirement.
B:This would validate a different case: An application seeks low-latency semantic recommendations but indexes product names as plain text without generated embeddings. It does not verify the outcome required here.
D:That corrective action is for a different case: A vector retrieval index stores embeddings from two incompatible model versions in one dimension-dependent collection. It does not resolve this incident.
Question 634
After resolving the following incident, which ongoing safeguard most directly prevents recurrence? A team changes an analytics library but releases a job without checking distributed task serialization compatibility.
- Include duplicate-event and retry fixtures in integration tests.
- Maintain supported language/runtime compatibility evidence in release reviews.
- Require DAG import and syntax checks in CI.
- Check service response enums when revising orchestration logic.
Explanation
Correct Answer
B:Maintain supported language/runtime compatibility evidence in release reviews. This safeguards against a recurring cause: Developer tests omit execution-context serialization and worker runtime dependencies.
Incorrect Answers
A:That safeguard addresses another case: An EventBridge rule matches both replayed and new events and starts two identical pipelines for the same batch. It is not the control for this recurring failure.
C:That safeguard addresses another case: An MWAA DAG was deployed but its Python import fails, so no scheduled run appears. It is not the control for this recurring failure.
D:That safeguard addresses another case: A workflow branches on a job-status string but overlooks the Cancelled terminal state, leaving executions stuck. It is not the control for this recurring failure.
Question 635
Two Kinesis consumers must independently recover after a deployment, but one team reuses the other's checkpoint store. Which engineering change best addresses the underlying issue?
- Choose a stable order identity and the correct aggregation or distinct-current-state logic.
- Use nullable-aware extraction and document sensible default or reject handling per field.
- Measure runtime and costs by job size and configure eligible autoscaling or differentiated job profiles.
- Separate durable per-application checkpoints and verify replay position semantics.
Explanation
Correct Answer
D:Separate durable per-application checkpoints and verify replay position semantics. It directly addresses the case cause (Consumer applications share state intended to represent separate progress positions.) and should be verified by: One consumer restarts without altering the other consumer's offset.
Incorrect Answers
A:That change addresses a nightly aggregation counts an order once per status-change event rather than once per order ID; it would not repair this incident.
B:That change addresses an API payload includes optional addresses but a transformation fails when a nullable nested object is absent; it would not repair this incident.
C:That change addresses a Glue job uses a fixed worker count at peak and wastes compute on tiny nightly delta batches; it would not repair this incident.
Question 636
A data transformation uses one Spark partition for a terabyte input and only one executor stays busy. Which technical dependency most directly explains this behavior?
- Orchestration does not deduplicate at the batch commit boundary.
- The scheduler cannot parse the DAG file and therefore cannot schedule its tasks.
- The processing layout prevents distributed parallelism.
- A wall-clock trigger is not a reliable data-completion dependency.
Explanation
Correct Answer
C:The processing layout prevents distributed parallelism. The corresponding fix is to repartition data with appropriate cardinality and avoid bottlenecking single-stage operations.
Incorrect Answers
A:This diagnosis addresses another AWS case: A production Glue workflow receives the same completed-file notification twice and creates two warehouse loads. The dependency here is different.
B:This diagnosis addresses another AWS case: An MWAA DAG was deployed but its Python import fails, so no scheduled run appears. The dependency here is different.
D:This diagnosis addresses another AWS case: An EventBridge schedule starts a warehouse load during an upstream S3 export rather than after all files are ready. The dependency here is different.
Question 637
A Glue transform joins events to reference data refreshed weekly, but old events are silently reclassified on every rerun. Which technical dependency most directly explains this behavior?
- The callback correlation contract is missing or insecure.
- Transient and permanent errors are not classified in the retry and catch policy.
- Event delivery and replay are not protected by a stable idempotency key.
- Reference data lacks effective-date or historical snapshot semantics.
Explanation
Correct Answer
D:Reference data lacks effective-date or historical snapshot semantics. The corresponding fix is to join by intended as-of validity and persist versioned reference mappings.
Incorrect Answers
A:This diagnosis addresses another AWS case: A Step Functions Standard workflow waits for an external vendor callback but the task token is lost in logs. The dependency here is different.
B:This diagnosis addresses another AWS case: A Step Functions state retries every failure including invalid schema, consuming budget without possibility of success. The dependency here is different.
C:This diagnosis addresses another AWS case: An EventBridge rule matches both replayed and new events and starts two identical pipelines for the same batch. The dependency here is different.
Question 638
A data quality gate compares raw and target record totals without accounting for intentional filter exclusions, falsely failing clean batches. Which technical dependency most directly explains this behavior?
- Changed query logic multiplies intermediate data and compute cost.
- The reconciliation rule ignores documented transformation semantics and accepted rejects.
- Pipeline health is not connected to operational alert thresholds.
- Skew creates straggler executors and prevents completion within the SLA.
Explanation
Correct Answer
B:The reconciliation rule ignores documented transformation semantics and accepted rejects. The corresponding fix is to reconcile accepted, filtered and quarantined records with explicit balance equations.
Incorrect Answers
A:This diagnosis addresses another AWS case: A Glue job begins consuming twice its normal DPU budget after a transformation adds a repeated cross join. The dependency here is different.
C:This diagnosis addresses another AWS case: A Glue job has failed for three mornings but no CloudWatch alarm exists, so dashboards show stale data. The dependency here is different.
D:This diagnosis addresses another AWS case: An ETL task in Glue times out only when a small fraction of keys dominate the daily partition. The dependency here is different.
Question 639
A data lake keeps years of detailed application logs in S3 Standard despite infrequent historical access and an approved 90-day hot tier. Which technical dependency most directly explains this behavior?
- Data placement increases redistribution for a frequently used join.
- Schema representation prevents typed spatial validation and query optimization.
- Storage lifecycle no longer reflects retention and access patterns.
- Unvalidated mixed-format source rows contaminated schema inference.
Explanation
Correct Answer
C:Storage lifecycle no longer reflects retention and access patterns. The corresponding fix is to apply reviewed S3 Lifecycle transition rules to suitable eligible storage classes.
Incorrect Answers
A:This diagnosis addresses another AWS case: A small dimension table joins huge Redshift facts, but an unnecessarily EVEN distribution causes repeated data transfer. The dependency here is different.
B:This diagnosis addresses another AWS case: A lake stores customer location coordinates as one unparsed text field while downstream joins require numeric latitude and longitude. The dependency here is different.
D:This diagnosis addresses another AWS case: A crawler changes a decimal sales column to string after encountering a nonnumeric banner row in one export. The dependency here is different.
Question 640
An Athena query for CloudTrail logs shows a missing day because the log data lake partitions were never registered. Which technical dependency most directly explains this behavior?
- Lake Formation grants and role identities are evaluated per workload principal.
- Audit log delivery and analytics catalog partition visibility are separate requirements.
- Cryptographic permissions exceed the data processing task's required keys.
- A service role cannot replace blocked VPC routes or security group rules.
Explanation
Correct Answer
B:Audit log delivery and analytics catalog partition visibility are separate requirements. The corresponding fix is to confirm CloudTrail delivery, then repair Athena catalog partitions or approved projection.
Incorrect Answers
A:This diagnosis addresses another AWS case: An Athena query against a Lake Formation resource succeeds for the analyst, yet a separate Glue ETL role cannot read that same table. The dependency here is different.
C:This diagnosis addresses another AWS case: A Glue data job decrypts source objects using a wildcard KMS permission spanning unrelated client keys. The dependency here is different.
D:This diagnosis addresses another AWS case: A Glue crawler in a VPC fails connection to a private RDS source even with the right IAM service role. The dependency here is different.
AWS Certified Data Engineer - Associate DEA-C01 Practice Tests