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 210
A team applies a reviewed fix for this situation: A monitoring rule sees one Athena query failure and pages the full team despite a known transient backend retry. Which observed outcome would most convincingly validate the fix?
- A transient injected failure remains logged, while persistent failures trigger on-call alerts.
- A forced timeout leads to one accepted logical batch and auditable retries.
- A maximum-size sample completes with a checkpointed durable output.
- Only a validated output is promoted to the consumer-facing location.
Explanation
Correct Answer
A:A transient injected failure remains logged, while persistent failures trigger on-call alerts. This verifies the result of correlate bounded retry recovery with sustained failure-rate or SLA breach conditions.
Incorrect Answers
B:This would verify another situation: A pipeline calls an AWS SDK to invoke Glue but drops the request token on network timeout and starts duplicate ETL runs. It does not validate the current case.
C:This would verify another situation: A monthly data load uses Lambda to copy hundreds of gigabytes in a single invocation and repeatedly times out. It does not validate the current case.
D:This would verify another situation: A Glue ETL job completes with warnings but the orchestrator assumes success despite a downstream quality check failing. It does not validate the current case.
Question 211
After resolving the following incident, which ongoing safeguard most directly prevents recurrence? A Glue PySpark job fails when an upstream field changes from integer to string in some JSON documents.
- Version the ingestion schema and alert on incompatible type drift.
- Audit DMS selection mappings after source schema changes.
- Associate build artifacts with immutable source hashes.
- Validate private-subnet routes and endpoint authorization in deployment tests.
Explanation
Correct Answer
A:Version the ingestion schema and alert on incompatible type drift. This safeguards against a recurring cause: Schema inference is unstable across source data evolution.
Incorrect Answers
B:That safeguard addresses another case: An on-premises Oracle source feeds DMS, but newly added tables never appear in a configured migration. It is not the control for this recurring failure.
C:That safeguard addresses another case: A SAM deployment packages an outdated Lambda handler because build artifacts from a previous branch were reused. It is not the control for this recurring failure.
D:That safeguard addresses another case: A batch import uses a Lambda in a private subnet that cannot reach S3 because no NAT or S3 endpoint path is configured. It is not the control for this recurring failure.
Question 212
A Glue JDBC extraction overloads a small OLTP database when hundreds of partitions open simultaneous connections. Which engineering change best addresses the underlying issue?
- Limit partitioned reads and schedule extraction within source performance budgets.
- Set supported concurrency/batch controls and scale throughput using measured sink limits.
- Update reviewed DMS table mapping and validate permissions and schema-change handling.
- Persist scoped task-token correlation securely and use supported callback completion.
Explanation
Correct Answer
A:Limit partitioned reads and schedule extraction within source performance budgets. It directly addresses the case cause (Extraction parallelism ignores source capacity and transaction impact.) and should be verified by: Replica or source CPU stays bounded while expected rows are extracted.
Incorrect Answers
B:That change addresses an SQS-backed processor overwhelms a downstream database after scaling consumers without any concurrency cap; it would not repair this incident.
C:That change addresses an on-premises Oracle source feeds DMS, but newly added tables never appear in a configured migration; it would not repair this incident.
D:That change addresses a Step Functions Standard workflow waits for an external vendor callback but the task token is lost in logs; it would not repair this incident.
Question 213
A scheduled Glue workflow triggers two jobs concurrently though the second requires the first output's partition. Which technical dependency most directly explains this behavior?
- Delivery buffering settings and destination behavior conflict with the required freshness.
- The event-to-job design starts a run per object without a batch completion boundary.
- Job ordering fails to express the producer-consumer dependency.
- Per-record interpreter overhead replaces optimized distributed expressions.
Explanation
Correct Answer
C:Job ordering fails to express the producer-consumer dependency. The corresponding fix is to add an explicit predecessor-success trigger or workflow state dependency.
Incorrect Answers
A:This diagnosis addresses another AWS case: A Firehose pipeline is expected to deliver individual records instantly, yet it buffers batches before S3 delivery. The dependency here is different.
B:This diagnosis addresses another AWS case: An S3 event notification triggers Glue for every file of a large export, producing hundreds of competing jobs. The dependency here is different.
D:This diagnosis addresses another AWS case: A PySpark transformation uses Python row-by-row loops to parse simple scalar fields and runs slowly. The dependency here is different.
Question 214
A Glue streaming job writes duplicate target rows when a microbatch is retried after a worker restart. Which engineering change best addresses the underlying issue?
- Use stable event identities and a sink merge/conditional-write design tied to checkpoint semantics.
- Use a stable ingestion event ID with conditional writes or a deduplicating transaction design.
- Set supported concurrency/batch controls and scale throughput using measured sink limits.
- Align source transaction-log retention, DMS checkpoints, and CDC task health with the recovery objective.
Explanation
Correct Answer
A:Use stable event identities and a sink merge/conditional-write design tied to checkpoint semantics. It directly addresses the case cause (The external sink operation is not idempotent across checkpoint replay.) and should be verified by: A forced retry leaves one intended row per business event.
Incorrect Answers
B:That change addresses an ingestion app writes identical DynamoDB item updates after a network retry and an audit counter increments twice; it would not repair this incident.
C:That change addresses an SQS-backed processor overwhelms a downstream database after scaling consumers without any concurrency cap; it would not repair this incident.
D:That change addresses a DMS CDC task finishes a full-load phase but stops capturing source table changes when the source transaction log expires; it would not repair this incident.
Question 215
An orchestrated data processing task receives tens of thousands of S3 events, causing Lambda reserved concurrency exhaustion. Select TWO complementary measures that correctly repair and validate this failure.
- Buffer events with a supported queue or stream and cap downstream concurrent workers.
- Remove all Lambda concurrency limits without evaluating downstream services.
- A shuffled fixture returns the same expected seven-day windows.
- A burst drains within the recovery SLA without widespread throttling.
- Correlate bounded retry recovery with sustained failure-rate or SLA breach conditions.
Explanation
Correct Answers
A:Buffer events with a supported queue or stream and cap downstream concurrent workers. The underlying cause is: The workflow has no buffering or explicit throughput regulation at the ingress boundary.
D:A burst drains within the recovery SLA without widespread throttling. This is the observable result required before accepting the correction.
Incorrect Answers
B:Remove all Lambda concurrency limits without evaluating downstream services. That shortcut conflicts with the stated requirement.
C:This would validate a different case: A data science notebook uses a rolling 7-day average without ordering by event date, so results change with input row order. It does not verify the outcome required here.
E:That corrective action is for a different case: A monitoring rule sees one Athena query failure and pages the full team despite a known transient backend retry. It does not resolve this incident.
Question 216
A pipeline sends database credentials in plaintext environment variables and writes them to CloudWatch on failure. Select TWO complementary measures that correctly repair and validate this failure.
- A controlled secret rotation works and logs contain no secret values.
- Store and rotate credentials in Secrets Manager with scoped runtime retrieval and redacted logs.
- Review CloudTrail denied access and compare effective role, bucket and KMS permissions.
- A controlled object cannot replicate to a disallowed Region, while authorized resilience works.
- Log the complete secret to aid postmortem troubleshooting.
Explanation
Correct Answers
A:A controlled secret rotation works and logs contain no secret values. This is the observable result required before accepting the correction.
B:Store and rotate credentials in Secrets Manager with scoped runtime retrieval and redacted logs. The underlying cause is: Secrets are handled as ordinary configuration and leaked through diagnostics.
Incorrect Answers
C:That corrective action is for a different case: An S3 lake read works with an administrator role but fails with the production ETL role, and the team assumes the dataset is corrupt. It does not resolve this incident.
D:This would validate a different case: A data lake in an EU Region replicates raw customer records to an unapproved Region for convenience. It does not verify the outcome required here.
E:Log the complete secret to aid postmortem troubleshooting. That shortcut conflicts with the stated requirement.
Question 217
An EMR Spark pipeline runs near memory limits because it caches multiple large datasets and never releases them. Which technical dependency most directly explains this behavior?
- Infrastructure drift makes deployments non-repeatable.
- Persisted dataset lifetimes exceed the stages that need them.
- The state machine lacks complete terminal-status handling.
- An unbounded driver collect breaks distributed processing assumptions.
Explanation
Correct Answer
B:Persisted dataset lifetimes exceed the stages that need them. The corresponding fix is to unpersist intermediate datasets and profile appropriate storage levels and partitions.
Incorrect Answers
A:This diagnosis addresses another AWS case: A serverless pipeline deploys with hand-edited IAM and S3 settings that differ among test and production accounts. The dependency here is different.
C:This diagnosis addresses another AWS case: A workflow branches on a job-status string but overlooks the Cancelled terminal state, leaving executions stuck. The dependency here is different.
D:This diagnosis addresses another AWS case: A distributed Spark process collects a large dataset to the driver and crashes from memory exhaustion. The dependency here is different.
Question 218
A DynamoDB table uses composite key `tenant#date` but most reads are by `transaction_id` only and become scans. Which engineering change best addresses the underlying issue?
- Restrict crawler targets, table prefixes and schema-change behavior to approved locations.
- Use a supported strongly consistent base-table or local-index read design where necessary.
- Use a supported ETL/write path to update the authoritative source, then refresh catalog metadata.
- Plan a suitable alternate index or redesign around known query requirements.
Explanation
Correct Answer
D:Plan a suitable alternate index or redesign around known query requirements. It directly addresses the case cause (Primary key design does not support a dominant lookup pattern.) and should be verified by: Transaction lookups are efficient without exposing cross-tenant records.
Incorrect Answers
A:That change addresses a Glue crawler crawls the entire lake for every hourly feed and adds unexpected tables from temporary staging prefixes; it would not repair this incident.
B:That change addresses a DynamoDB application requires strongly consistent reads immediately after a successful critical write but queries a global secondary index; it would not repair this incident.
C:That change addresses a Redshift query joins external and local datasets but expects Spectrum to automatically overwrite remote S3 data during UPDATE; it would not repair this incident.
Question 219
A nightly data product uses a moving seven-day window but incoming source events arrive late and are excluded without a watermark policy. Select TWO complementary measures that correctly repair and validate this failure.
- Athena reads only successfully committed transform output after the Glue run ends.
- A known delayed event updates the intended window according to policy.
- Normalize event time and use the required explicit UTC or business timezone boundaries.
- Implement event-time windows with a supported lateness allowance and correction/backfill strategy.
- Discard all records that arrive after midnight regardless of event time.
Explanation
Correct Answers
B:A known delayed event updates the intended window according to policy. This is the observable result required before accepting the correction.
D:Implement event-time windows with a supported lateness allowance and correction/backfill strategy. The underlying cause is: Data freshness and event-time completeness assumptions are not explicit.
Incorrect Answers
A:This would validate a different case: A Step Functions data pipeline invokes Glue using a request/response integration and immediately launches Athena against unfinished output. It does not verify the outcome required here.
C:That corrective action is for a different case: A daily sales pivot groups event timestamps by the viewer's local date and differs from the regulatory UTC day. It does not resolve this incident.
E:Discard all records that arrive after midnight regardless of event time. That shortcut conflicts with the stated requirement.
Question 220
A distributed Spark process collects a large dataset to the driver and crashes from memory exhaustion. Which technical dependency most directly explains this behavior?
- The transformation changes record granularity without an explicit explode/join design.
- Retry logic does not isolate staging and atomic commit boundaries.
- An unbounded driver collect breaks distributed processing assumptions.
- The external sink operation is not idempotent across checkpoint replay.
Explanation
Correct Answer
C:An unbounded driver collect breaks distributed processing assumptions. The corresponding fix is to keep transformations distributed and collect only bounded diagnostic samples.
Incorrect Answers
A:This diagnosis addresses another AWS case: A Glue job reading nested JSON flattens arrays incorrectly and duplicates parent order totals. The dependency here is different.
B:This diagnosis addresses another AWS case: A Glue workflow reruns yesterday's partition after a partially failed daily job but overwrites already-approved historical results. The dependency here is different.
D:This diagnosis addresses another AWS case: A Glue streaming job writes duplicate target rows when a microbatch is retried after a worker restart. The dependency here is different.
AWS Certified Data Engineer - Associate DEA-C01 Practice Tests