AWS DOP-C02Trap Reference

Commonly Confused Services on DOP-C02

DOP-C02 is largely about deployment judgment, operational complexity, and choosing the right AWS-native mechanism without overbuilding. Service disambiguation is one part of that — this page covers the tool pairs in the CI/CD and automation stacks where wrong defaults cost points.

Each section below gives you the deciding signal, a quick check, and why the wrong answer keeps looking right.

AWS CodeBuildAWS CodeDeployAWS CodePipeline
#1

Build vs. deploy vs. orchestrate

All three have "Code" in the name and appear together in CI/CD questions, so candidates conflate their specific roles.

Deciding signal

CodeBuild compiles source code, runs unit tests, and produces build artifacts — it is the CI build stage. CodeDeploy takes a deployable artifact (from S3 or an ECR image) and deploys it to EC2 instances, ECS services, Lambda functions, or on-premises servers using a configurable strategy. CodePipeline orchestrates the stages: source → build → test → deploy → approval. A question asking "which service executes the rolling deployment to EC2" points to CodeDeploy. A question asking "which service triggers the build when a commit is pushed and then deploys if tests pass" points to CodePipeline as the orchestrator. DOP-C02 specifically tests whether candidates can distinguish the deployment executor (CodeDeploy) from the pipeline orchestrator (CodePipeline).

Quick check

Is the requirement to compile and test source (CodeBuild), execute a deployment strategy on compute targets (CodeDeploy), or orchestrate the full CI/CD workflow end-to-end (CodePipeline)?

Why it looks right

CodePipeline is associated with CI/CD broadly and candidates apply it to deployment-specific questions. CodeDeploy is the service that actually controls the deployment strategy — blue/green, canary, rolling.

AWS CloudFormationAWS CDKAWS SAM
#2

Declarative YAML/JSON IaC vs. programmatic IaC vs. serverless application model

All three provision AWS infrastructure from code, so candidates treat them as equivalent IaC tools.

Deciding signal

CloudFormation is the foundation — it accepts declarative YAML or JSON templates and provisions resources. CDK (Cloud Development Kit) lets you define infrastructure in TypeScript, Python, Java, or C#; it synthesizes to CloudFormation templates. SAM (Serverless Application Model) is a CloudFormation extension with shorthand syntax for Lambda functions, API Gateway, DynamoDB, and Step Functions — it simplifies serverless resource definitions and includes the SAM CLI for local testing. The signal is the abstraction level: raw YAML/JSON (CloudFormation), code in a general-purpose language (CDK), or shorthand for serverless resources (SAM). DOP-C02 tests which tool is appropriate for a given developer workflow.

Quick check

Is this raw declarative YAML/JSON templates (CloudFormation), infrastructure defined in a programming language (CDK), or serverless resources with shorthand and local testing via CLI (SAM)?

Why it looks right

CDK is the modern high-abstraction answer and candidates default to it. SAM is correct when the scenario specifically involves serverless resources and local Lambda testing — and CloudFormation is correct when the scenario emphasizes declarative templates without code generation.

CodeDeploy In-PlaceCodeDeploy Blue/GreenCodeDeploy CanaryCodeDeploy Linear
#3

Replace in place vs. swap environments vs. gradual percentage shift vs. equal increment shift

All four are CodeDeploy deployment strategies and candidates associate blue/green with all "zero-downtime" scenarios.

Deciding signal

In-Place deploys to existing instances — stops the current app, deploys the new version, restarts. There is downtime. Blue/Green provisions a new set of instances with the new version, validates them, then shifts traffic from old (blue) to new (green). Old instances remain for rollback. Canary shifts a small percentage of traffic (e.g., 10%) to the new version, waits, then shifts the rest if the first segment succeeds. Linear shifts traffic in equal increments on a fixed schedule (e.g., 10% every 5 minutes). Canary is two steps; Linear is many equal steps. When the scenario describes validating with a small subset of traffic first before full cutover, Canary. When it describes a gradual equal-increment shift, Linear.

Quick check

Is this replacing instances directly with downtime (In-Place), swapping two full environments (Blue/Green), a two-step shift to test a small percentage first (Canary), or equal incremental shifts over time (Linear)?

Why it looks right

Blue/Green is the default "zero-downtime deployment" answer. Canary is correct when the scenario emphasizes testing with a small traffic percentage before a full shift — a different risk model from Blue/Green.

AWS Elastic BeanstalkAWS CodeDeployAWS OpsWorks
#4

PaaS application deployment vs. deployment strategy execution vs. Chef/Puppet configuration management

All three deploy and configure applications on EC2, so candidates blur them in operations questions.

Deciding signal

Elastic Beanstalk is a PaaS: you upload application code (ZIP or Docker) and Beanstalk provisions EC2, ELB, Auto Scaling, and CloudWatch automatically. It manages the infrastructure lifecycle. CodeDeploy is a deployment service that takes a pre-built artifact and deploys it to target compute using a defined strategy — it requires the infrastructure to already exist. OpsWorks manages Chef or Puppet configuration management: you define layers and recipes, and OpsWorks handles software configuration on instances. The signal is the abstraction level: code upload with infrastructure provisioning (Beanstalk), deployment to existing infrastructure with strategy control (CodeDeploy), or Chef/Puppet configuration management (OpsWorks).

Quick check

Is this uploading application code with AWS managing all infrastructure (Beanstalk), deploying a build artifact to existing infrastructure with a defined strategy (CodeDeploy), or managing server configuration with Chef or Puppet recipes (OpsWorks)?

Why it looks right

Elastic Beanstalk and CodeDeploy both "deploy applications" and candidates conflate them. Beanstalk provisions infrastructure; CodeDeploy requires existing infrastructure and controls the deployment strategy.

AWS Systems Manager Parameter StoreAWS Secrets Manager
#5

Configuration and non-rotating secrets vs. credentials with automatic rotation

Both store sensitive values securely and both appear in "store credentials" questions.

Deciding signal

Parameter Store stores strings, string lists, and encrypted secure strings via KMS. Standard parameters are free; there is no rotation mechanism built in. Secrets Manager stores secrets with a built-in rotation engine: it natively integrates with RDS, Redshift, and DocumentDB to rotate passwords on a defined schedule, and supports Lambda-based rotation for any other service. It charges per secret per month. On DOP-C02, Secrets Manager is specifically the answer when the scenario mentions automatic credential rotation or when the secret is a database password that must be rotated without application changes. Parameter Store is the answer for non-rotating configuration, environment variables, or per-parameter storage at scale where cost matters.

Quick check

Does the secret need automatic rotation on a schedule (Secrets Manager), or is this configuration or a non-rotating secret stored at scale (Parameter Store)?

Why it looks right

Secrets Manager sounds more appropriate for all secrets, but Parameter Store is the cost-effective correct answer when automatic rotation is not required. The exam specifically tests awareness of this distinction.

Amazon CloudWatchAWS X-RayAWS CloudTrail
#6

Operational metrics and alarms vs. distributed request tracing vs. API audit

All three are observability services and "monitoring" in a question does not distinguish them.

Deciding signal

CloudWatch collects metrics, ingests logs, triggers alarms, and drives dashboards — it answers "how is this resource performing?" X-Ray traces requests through distributed systems: it shows which service called which, how long each segment took, and where errors occurred across Lambda, API Gateway, ECS, and other instrumented services. It answers "where is this request failing or slowing down across my microservices?" CloudTrail records API calls — who made them, when, from where — for account-level audit and governance. When a scenario describes debugging latency in a Lambda-to-DynamoDB request chain, X-Ray. When it describes alerting on error rate, CloudWatch. When it describes who deleted a deployment configuration, CloudTrail.

Quick check

Is this metrics and alarms on resource performance (CloudWatch), tracing request paths across distributed services (X-Ray), or auditing who made which API calls (CloudTrail)?

Why it looks right

CloudWatch is the default observability answer. X-Ray is specifically for distributed tracing across service boundaries — a problem that CloudWatch metrics alone cannot diagnose.

AWS CodeArtifactAmazon ECRAmazon S3
#7

Package registry vs. container image registry vs. object storage for build artifacts

All three store build outputs, so candidates default to S3 for all artifact storage scenarios.

Deciding signal

S3 is generic object storage — CodePipeline and CodeBuild use it to pass build artifacts between pipeline stages. It stores ZIP files, binaries, and other build outputs. ECR is a fully managed container image registry — it stores, versions, and distributes Docker and OCI images. It is the specific answer when the build output is a container image. CodeArtifact is a managed artifact repository for software packages — Maven, npm, pip, NuGet, and others. It proxies public registries, caches packages, and stores your internal private packages. When the scenario involves sharing npm or Maven packages internally, CodeArtifact. Container images, ECR. Generic pipeline artifacts between stages, S3.

Quick check

Is this storing container images (ECR), caching or sharing npm/Maven/pip packages (CodeArtifact), or passing generic build artifacts between pipeline stages (S3)?

Why it looks right

S3 is the familiar artifact storage answer. CodeArtifact is specifically correct when the scenario involves software package dependencies — candidates often miss the distinction between a package registry and generic file storage.

CloudFormation StackSetsCloudFormation Stacks
#8

Multi-account multi-region deployment vs. single-account resource provisioning

Both deploy CloudFormation templates, so candidates treat StackSets as a scaling enhancement of Stacks without checking the scope.

Deciding signal

CloudFormation Stacks deploy resources in a single AWS account and region. StackSets extend this to deploy the same template across multiple accounts and regions simultaneously — managed centrally from an administrator account. StackSets require an Organizations setup for service-managed mode or explicit delegated admin configuration for self-managed mode. When the scenario describes deploying a resource configuration consistently across all accounts in an organization (e.g., a CloudTrail configuration, a security baseline), StackSets. When the scenario involves managing infrastructure within a single account, Stacks.

Quick check

Is this deploying resources in one account and region (Stacks), or deploying the same template across multiple accounts and regions centrally (StackSets)?

Why it looks right

StackSets sound like a feature addition that all deployments could use. Candidates overlook them in multi-account scenarios, defaulting to Stacks and not realizing StackSets is the specific answer for organizational deployment.

AWS Systems Manager AutomationAmazon EventBridgeAWS Lambda
#9

Runbook-based operational actions vs. event routing vs. custom compute logic

All three trigger automated actions in response to events, so candidates pick Lambda as the default automation mechanism.

Deciding signal

Lambda runs custom code — it is the right answer when the automation requires logic beyond predefined runbooks. Systems Manager Automation runs pre-built or custom runbooks (SSM Documents) against AWS resources — restarting instances, patching AMIs, creating snapshots, taking resource compliance actions. It integrates with Maintenance Windows and provides approval steps and rate controls. EventBridge routes events between services based on rules — it is the event bus, not the execution engine. The typical pattern is EventBridge detects an event and triggers an SSM Automation runbook or Lambda function. When the scenario describes executing an operational action on an EC2 instance or resource using a predefined runbook, SSM Automation. When it describes custom logic, Lambda.

Quick check

Is this executing a predefined operational runbook against AWS resources (SSM Automation), routing an event to a target based on rules (EventBridge), or running custom application logic (Lambda)?

Why it looks right

Lambda is the default automation answer because it can do anything. SSM Automation is correct when the scenario describes executing a defined operational action — particularly one that targets EC2 instances or follows an approval workflow.

AWS OrganizationsAWS Control TowerAWS Config
#10

Account structure and SCPs vs. landing zone automation vs. resource compliance rules

All three enforce governance across AWS environments, so candidates blend them in multi-account compliance questions.

Deciding signal

Organizations is the structural foundation: account grouping into OUs and service control policies (SCPs) that restrict what any account can do. It does not provision or baseline accounts. Control Tower automates the creation of a well-governed multi-account landing zone — it provisions accounts with pre-configured security baselines, logging, and guardrails using Organizations under the hood. Config evaluates individual resource configurations against rules and tracks configuration history. DOP-C02 tests whether candidates can separate account-level permission guardrails (Organizations SCPs), automated account provisioning and baseline governance (Control Tower), and continuous resource-level compliance evaluation (Config). These three often work together but each answers a different governance question.

Quick check

Is this restricting what any account can do (Organizations SCPs), automating a governed multi-account landing zone with account provisioning (Control Tower), or continuously evaluating and enforcing specific resource configurations (Config)?

Why it looks right

Organizations and Control Tower are closely related and candidates conflate them. Control Tower automates what would otherwise require manual Organizations configuration — the distinction matters when the scenario describes setting up new accounts with baseline compliance rather than managing existing SCP policy.

Train these confusions, not just read them

10 DOP-C02 questions. Pattern-tagged with trap analysis. Free, no signup required.

Start DOP-C02 Mini-Trainer →