Free DVA-C03 sample questions
These five scenario questions are original and cover services that matter for the Developer - Associate exam in both its outgoing and updated versions. They are not taken from any exam. Try each before opening the answer.
Question 1. An order-processing Lambda function reads messages from an SQS standard queue. Occasionally a message is delivered twice and a customer is charged twice. What is the best way to prevent duplicate charges?
- A. Increase the function timeout to the maximum
- B. Make the handler idempotent by recording the order ID with a conditional write before charging
- C. Increase the batch size so each message is processed in a larger group
- D. Disable retries on the event source mapping
Show answer
Answer: B
Standard queues deliver at least once, so a message can arrive more than once. Making the handler idempotent, for example with a conditional write keyed on the order ID, makes repeated delivery harmless. A longer timeout does not stop redelivery, raising batch size changes throughput not correctness, and disabling retries risks losing messages.
Want more questions like this? Full DVA-C03 practice tests →
Question 2. A DynamoDB table stores one item per sensor reading with the partition key set to the data-centre name. Only three data centres exist and writes are throttled although the table has plenty of provisioned capacity overall. What is the most likely cause and the best fix?
- A. Items are too large; compress the attributes
- B. Switch the table to a customer managed KMS key
- C. Add a global secondary index on the data-centre name
- D. The partition key has low cardinality; use a higher-cardinality key such as the sensor ID
Show answer
Answer: D
With only three partition key values, traffic concentrates on a few partitions, creating hot partitions and throttling despite spare total capacity. A key with high cardinality, such as a sensor ID combined with the data centre, spreads the load. Larger items, encryption settings and a global secondary index on the same keys do not fix key distribution.
Want more questions like this? Full DVA-C03 practice tests →
Question 3. An application running on Amazon ECS must read objects from a single S3 bucket. Which approach follows least privilege and avoids long-lived credentials?
- A. Attach a task IAM role whose policy allows only the required S3 actions on that bucket
- B. Store an IAM user access key in a container environment variable
- C. Bake an access key into the container image
- D. Attach an administrator policy to the task role for simplicity
Show answer
Answer: A
A task role gives the containers temporary credentials, and scoping its policy to the one bucket and the needed actions follows least privilege. Access keys in environment variables or an image are long-lived secrets, and an administrator role is far broader than required.
Want more questions like this? Full DVA-C03 practice tests →
Question 4. A team must encrypt a 2 GB file client-side before uploading it to S3 using AWS KMS. Why can the KMS Encrypt API not be used on the whole file directly, and what is the standard approach?
- A. Compress the file first, then call Encrypt once
- B. Split the file into 4 KB parts and call Encrypt on each part
- C. Use envelope encryption: generate a data key with KMS, encrypt the file locally with it, and store the encrypted data key with the file
- D. Make the bucket public so encryption is not required
Show answer
Answer: C
The KMS Encrypt operation handles only small payloads of up to 4 KB, so large data is encrypted with a data key generated by KMS (envelope encryption), and the data key is stored encrypted alongside the data. Compression, splitting into 4 KB calls, or using a public bucket do not match the standard design.
Want more questions like this? Full DVA-C03 practice tests →
Question 5. Users report that an API backed by API Gateway, Lambda and DynamoDB is slow only for some requests. The team needs to see which downstream call adds the latency for individual requests. What should they use?
- A. CloudWatch billing metrics
- B. Enable AWS X-Ray tracing and inspect the service map and traces
- C. An IAM credential report
- D. An S3 inventory report
Show answer
Answer: B
AWS X-Ray traces requests across services and shows segments and subsegments with timing, so the slow downstream call is visible per request. Aggregate metrics show averages but not the per-request path, and billing reports or IAM credential reports are unrelated to latency.
Want more questions like this? Full DVA-C03 practice tests →
What to take from these
Each scenario turns on a constraint: at-least-once delivery, key cardinality, credentials that must not be stored, a size limit, and per-request visibility. Developer questions usually test whether you know how a service behaves under that constraint, not a definition. If these felt easy, move to timed practice in week 4 of the study plan. If the security items were weaker, revisit AWS IAM basics.