Amazon Web Services · DEA-C01

DEA-C01 Exam: Complete AWS Data Engineer Guide

Current DEA-C01 exam guide covering the $150 USD fee, 130-minute format, 720 passing score, four domains, AWS data services, and preparation.

Last verified September 6, 2026

DEA-C01 exam at a glance

Facts last verified September 6, 2026

Domain breakdown

Who should take the DEA-C01 exam?

DEA-C01 is for data engineers and data architects who build reliable data platforms on AWS. The role covers ingestion, transformation, orchestration, storage, modeling, operations, security, and governance. It is not limited to writing ETL code or operating one analytics service.

AWS describes an ideal candidate with two to three years of data engineering or architecture experience and one to two years of hands-on AWS work. You should understand data volume, velocity, variety, schema, quality, privacy, and operational requirements and translate them into AWS service and design decisions.

There is no certification prerequisite. Candidates with general AWS architecture knowledge still need data-specific practice: partitioning, file formats, streaming semantics, job orchestration, catalogs, access patterns, lineage, monitoring, and recovery.

Skills measured on DEA-C01

The official DEA-C01 exam guide defines four domains. Ingestion and transformation plus store management account for 60% of scored content, but a design is incomplete unless it can be secured, observed, and recovered.

Data Ingestion and Transformation (34%)

Choose ingestion patterns based on source, throughput, latency, ordering, replay, error handling, and consumer requirements. Batch transfer can use services such as AWS DataSync, Transfer Family, Database Migration Service, or scheduled jobs. Streaming solutions may use Kinesis Data Streams, Kinesis Data Firehose, Amazon Managed Streaming for Apache Kafka, or service-native event sources.

Transformation choices include AWS Glue ETL, Glue DataBrew, Amazon EMR, Lambda, SQL engines, and stream processing. Know when serverless managed processing is sufficient and when a distributed Spark or Hadoop environment is justified. Understand file formats such as Parquet, ORC, JSON, and CSV, along with compression, partitioning, and small-file tradeoffs.

Pipeline orchestration may involve Step Functions, Managed Workflows for Apache Airflow, Glue workflows, EventBridge, and service integrations. Design idempotent tasks, retries with backoff, dead-letter handling, checkpoints, dependency controls, and restartability. A successful job that duplicates records during retry is not a reliable pipeline.

Programming concepts include data manipulation, APIs, SQL, and infrastructure automation. Practice recognizing when pushdown predicates, partition pruning, parallelism, or incremental processing reduce time and cost.

Data Store Management (26%)

Select stores from access patterns rather than familiarity. S3 is durable object storage and a common data-lake foundation; Redshift serves analytical warehousing; RDS and Aurora serve relational transactions; DynamoDB supports key-value and document access at scale; OpenSearch supports search and analytics; and purpose-built stores solve graph, time-series, or caching requirements.

The Glue Data Catalog and Lake Formation support discovery and governed access. Understand crawlers, tables, partitions, schema inference, cross-account sharing, and fine-grained permissions. Catalogs describe data; they do not by themselves guarantee that underlying data is correct or accessible.

Lifecycle design covers S3 storage classes and policies, retention, deletion, backup, archival, replication, and legal requirements. Balance recovery objectives against cost. Know how versioning, Object Lock, and replication address different risks.

Data modeling includes normalized and denormalized designs, star schemas, fact and dimension tables, slowly changing dimensions, keys, indexes, distribution, sort strategies, and DynamoDB single-table patterns. Plan for schema evolution and compatibility so producers can change without silently breaking consumers.

Data Operations and Support (22%)

Automate infrastructure and pipeline changes with CloudFormation or other supported infrastructure- as-code tools. Use CI/CD, version control, test data, quality checks, and separate environments. Data quality rules should detect validity, completeness, uniqueness, timeliness, and referential problems before bad data reaches downstream decisions.

Monitoring spans CloudWatch metrics, logs, alarms, dashboards, EventBridge events, CloudTrail, and service-specific history. Establish baselines for freshness, throughput, lag, error rate, job duration, and cost. A green compute metric is insufficient if business data arrived late.

Troubleshooting begins by locating the failing layer: permissions, networking, source availability, schema, capacity, code, partition layout, or downstream throttling. Know how to retry safely, reprocess from a checkpoint, quarantine bad records, restore data, and communicate data incidents.

Data Security and Governance (18%)

Apply least privilege with IAM roles, resource policies, Lake Formation permissions, and cross- account controls. Prefer temporary credentials and service roles over embedded access keys. Understand how identity policies, resource policies, permissions boundaries, and organization controls interact.

Protect data in transit and at rest using TLS and encryption services such as KMS. Key policies, grants, rotation, and cross-account access can determine whether an otherwise correct pipeline works. Secrets Manager and Systems Manager Parameter Store help keep database credentials and configuration out of code.

Governance covers classification, discovery, lineage, quality, retention, audit, and privacy. Services such as Macie can discover sensitive S3 data, CloudTrail records API activity, and AWS Config evaluates configurations. Masking, tokenization, row or column filtering, and data minimization may be necessary depending on use and regulation.

How to prepare for DEA-C01

Build one batch pipeline and one streaming pipeline. For batch, land source files in S3, catalog them, transform to columnar format, partition them, query them, and implement quality checks. For streaming, ingest events, handle malformed records, monitor lag, and make the consumer idempotent.

Then introduce failures: revoke a KMS permission, change a schema, create many small files, throttle a destination, or fail midway through a job. Diagnose from logs and metrics and recover without duplicating or losing data. Document RPO, RTO, ownership, and alert thresholds.

Map each official objective to a real command, console workflow, or design decision. Use AWS's official practice questions to find blind spots. Spend study time by domain weight, while continuing to solve cross-domain scenarios.

DEA-C01 practice questions

Good practice questions include data volume, latency, ordering, transformation, access, and cost constraints. State why the selected service and configuration meet those constraints and why a plausible alternative fails. This develops architecture judgment instead of keyword matching.

Use objective-aligned practice only. Unauthorized dumps violate AWS rules and cannot teach the operational reasoning needed when a pipeline fails.

DEA-C01 compared with other AWS exams

Solutions Architect – Associate covers broad architecture; DEA-C01 goes deeper into ingestion, storage, transformation, modeling, and data operations. Machine Learning Engineer – Associate focuses on ML workloads and depends on well-engineered data, while Security – Specialty goes deeper into controls and investigations. Choose DEA-C01 when data platforms are the core job.

Career value of DEA-C01

The credential aligns with data engineer, analytics engineer, data platform engineer, and data architect roles using AWS. It is strongest when supported by a portfolio that shows tested pipelines, data contracts, governance, monitoring, and cost-aware design.

Certification cannot replace data fundamentals. Employers benefit when you can explain not only which service you chose, but how the system behaves under schema changes, retries, growth, and security incidents.

Exam-day notes

• You have 130 minutes for 65 questions.
• Fifty questions are scored and 15 are unscored and unidentified.
• The US fee is $150 before taxes or regional adjustments.
• The minimum scaled score is 720, and domain performance is compensatory.
• Answer every question because AWS applies no guessing penalty.

Prep resources

← More Amazon Web Services exams