DVA-C02 study plan

Your step-by-step plan

1. Map the four domains to your real experience

   Day 1

   Start from the exam guide's four domains and their weights: Development with AWS Services is the largest, then Security, Deployment, and Troubleshooting and Optimization. Mark which you already do at work and which you only read about. Most developers are strong on writing code and weak on the deployment tooling and observability services, so plan your time against the gaps, not the comfort zones.

2. Drill the core build services until the APIs are reflexive

   Week 1

   Spend the most time on Lambda, DynamoDB, and the messaging trio. Write a Lambda handler, set its memory, timeout, reserved and provisioned concurrency, and wire an event source mapping with a dead-letter queue. Model a DynamoDB table, run a query and a scan, add a global secondary index, and do a conditional write. Send a message through SQS, fan out with SNS, and route with EventBridge. The exam tests these constantly.

3. Learn identity, encryption, and secrets as decision rules

   Week 2

   Practise the security discriminations until they are automatic: an assumed IAM role with AWS STS over stored keys, a Cognito user pool versus identity pool, ID token versus access token, Secrets Manager versus Parameter Store, AWS managed versus customer managed KMS keys. For each, write the one-line reason it wins so you can answer in seconds under exam pressure.

4. Build and ship with the deployment tooling

   Week 3

   Define a small application with AWS SAM, deploy it with a change set, and publish a Lambda version behind an alias. Set up a CodePipeline with a CodeBuild stage and a manual approval. Do a CodeDeploy canary and a blue/green cutover so the strategy differences are muscle memory, not trivia. Learn how API Gateway stages and stage variables promote the same deployment across environments.

5. Instrument, trace, and tune

   Week 4

   Turn on X-Ray for a multi-service request and read the service map. Run a CloudWatch Logs Insights query over structured JSON logs. Emit a custom metric with the embedded metric format and alarm on it through SNS. Tune a Lambda function's memory to see the vCPU and cost effect, and add provisioned concurrency to kill cold starts. This domain is small but easy marks once the tool-to-symptom mapping is fixed.

6. Practise on scenario questions and review every wrong option

   Week 5

   Move to full scenario questions and treat each one as a drill: name the requirement in the stem, predict the answer, then read why every distractor is wrong. The adjacent-service traps (SQS versus SNS, query versus scan, ID versus access token, canary versus blue/green) are where marks are won and lost. Keep a list of the discriminations you miss and re-drill them.

7. Simulate the exam and close the gaps

   Week 6

   Sit timed sets of about 65 questions in 130 minutes so the pace is familiar. Flag-and-return on long scenarios rather than stalling. After each set, sort misses by domain and task statement, go back to the service or API behaviour behind each, and re-drill. Stop when you are consistently scoring comfortably above the pass mark on unseen questions.

Frequently asked questions