Home › Certifications › AWS Data Engineer Associate › Data Operations and Support
33 multiple-choice questions and 19 flashcards on Data Operations and Support, about 22% of the AWS Data Engineer Associate bank. Every one carries a written rationale.
Data Operations and Support is one of 4 chapters in CoStudy's AWS Data Engineer Associate bank, and it holds 33 of the bank's 150 multiple-choice questions — roughly 22% of the total. That proportion is not arbitrary: chapters follow the certifying body's published exam outline, and the number of questions in each is set by that domain's published weight, so the share of your practice time this chapter takes matches the share of the real exam it accounts for.
Studying by chapter is worth doing once you have a diagnostic score. A single overall percentage tells you whether you are close; it does not tell you which domain is dragging. Working a weak chapter in isolation, and re-testing it in isolation, is the fastest way to move a score that has stalled — and it is why the mock exams in CoStudy report by domain rather than as one number.
10 questions drawn from this chapter, with the full rationale shown — the controlling principle behind the right answer, and why each wrong option tempts and fails.
Which is the FIRST place to check when an Athena query returns HIVE_CURSOR_ERROR on a Parquet table?
Answer: B — The underlying Parquet files for corruption or a schema mismatch with the Catalog
A) Budget alarms produce different errors; HIVE_CURSOR_ERROR is a data-side issue. B) HIVE_CURSOR_ERROR usually points to corrupt files or a schema mismatch between the Parquet footer and the Catalog table. C) Permissions surface as AccessDenied. D) Versioning does not cause HIVE_CURSOR_ERROR.
Which is the MOST cost-effective way to run a nightly 200 GB Spark transformation?
Answer: D — EMR Serverless with capacity limits and auto-stop after job completion
A) A 24-hour cluster for a nightly job wastes idle hours. D) EMR Serverless auto-scales and stops after the job — pay only for actual runtime. C) UDFs cannot replicate arbitrary Spark logic and consume Redshift resources. B) Athena CTAS can work but Spark transformations often require imperative code Athena SQL cannot express.
Which sequence describes a Step Functions Wait-For-Callback pattern?
Answer: B — Task publishes a token and pauses; an external system returns the token to resume
A) That is a polling pattern, not callback. B) The Task With Callback pattern uses .waitForTaskToken — the workflow pauses until an external process calls SendTaskSuccess with the token. C) That is a fire-and-forget with periodic re-check. D) That is synchronous integration.
Which is TRUE about the Glue Workflow feature?
Answer: C — Workflows compose crawlers and jobs into a DAG with triggers between them
C) Glue Workflows link crawlers, jobs, and triggers into a DAG scoped to Glue components. B) Workflows are Glue-only; Step Functions is broader. A) Workflows do not run on AWS Batch. D) MWAA is not required.
Which is TRUE about the Redshift workload manager (WLM)?
Answer: B — WLM assigns queries to queues with concurrency limits and memory allocations
B) WLM (or Auto WLM) routes queries into queues by predicates and enforces concurrency and memory. A) WLM has nothing to do with Athena conversion. C) WLM is not a network layer. D) WLM does not replace IAM.
Which service BEST orchestrates a multi-stage ETL DAG using Apache Airflow?
Answer: D — Amazon Managed Workflows for Apache Airflow (MWAA) hosting DAGs on S3
D) MWAA is managed Airflow — DAGs on S3, workers auto-scale, metadata DB provided. B) Step Functions is a state machine, not Airflow. C) EventBridge Scheduler runs jobs on time, not a DAG. A) Self-managed EC2 misses managed operational benefits.
Which is TRUE about EMR cluster instance fleets?
Answer: B — Fleets mix instance types and capacities to satisfy target vCPU and memory needs
B) Instance fleets let EMR blend types to hit target capacity — improving Spot availability. A) Fleets are heterogeneous by design. C) Fleets are available in classic EMR. D) Fleets support on-demand and Spot mixed.
Which is the BEST way to alert on a failed Glue job?
Answer: A — CloudWatch Events (EventBridge) rule matching Glue job state=FAILED to SNS topic
A) EventBridge natively emits Glue job state-change events; SNS delivers alerts. B) Polling is wasteful. C) Drift detects config change, not runtime failure. D) Screen scraping is not a real solution.
Which is a REAL cause of Redshift skew in a KEY-distributed table?
Answer: C — Most rows share the same distribution key value, sending them all to one slice
C) Skew happens when one distribution-key value dominates rows; that value's slice becomes hot. B) Sort and distribution keys are orthogonal choices. A) ATO helps balance distribution over time. D) Concurrency scaling adds compute, not skew.
Which Step Functions integration BEST processes millions of files in parallel?
Answer: A — A Map state in distributed mode reading a JSON list from S3
B) A serial for-loop bottlenecks the whole workflow. A) Distributed Map fans out over items with high parallelism and reads inputs from S3 without loading the whole list in state. C) Choice picks branches, not parallelism. D) Wait states add latency, not throughput.
4 cards from the 19 in this chapter.
Passing score?
720/1000.
Managed Workflows for Apache Airflow (MWAA)?
Managed Apache Airflow service. For complex orchestration with DAGs in Python.
Glue Auto Scaling — what does it do?
Dynamically adjusts DPU count during a job based on workload, reducing cost on small data and adding capacity on large data.
Glue Data Quality?
Apply DQ rules to S3 datasets + Glue jobs. Detect issues, calculate scores, integrate with workflows.
These are a sample. The full Data Operations and Support chapter runs 52 items with per-chapter progress tracking, on the web and in the iOS app.
Open AWS Data Engineer Associate in CoStudy →