Free AWS Certified Developer - Associate practice — 6 questions on Development with AWS Services, with explanations. No sign-up.
Full 12-question mixed test →
Question 1 of 6 · Development with AWS Services
A DynamoDB table stores inventory counts. Two Lambda functions running concurrently both read an item's 'stock' attribute and 'version' attribute, then attempt to decrement stock and increment version. Without any coordination, one update can silently overwrite the other, causing incorrect stock counts. Which approach BEST prevents this lost-update problem?
This is optimistic locking: read the version, then write only if the version hasn't changed (ConditionExpression version = :expectedVersion), incrementing version on success. A ConditionalCheckFailedException tells the caller to retry with fresh data, preventing lost updates.
Question 2 of 6 · Development with AWS Services
An order-processing platform publishes order events containing deeply nested JSON with fields like order.items[].category and order.customer.tier. Different consuming services need to react only when specific nested field values match certain patterns, and some rules must also fire on a recurring schedule independent of incoming events. Which service should be used to route and filter these events?
EventBridge event patterns can match deeply nested JSON fields (not just top-level message attributes), route to multiple targets, and EventBridge also natively supports scheduled (cron/rate) rules on the same bus, satisfying both requirements in one service.
Question 3 of 6 · Development with AWS Services
A Lambda function backing a mobile app has reserved concurrency set to 50 to protect a downstream RDS database. During a marketing push, traffic spikes sharply at a known time each morning and users report multi-second delays during the first minute of the spike, even though total concurrent invocations stay under 50. What should be configured to eliminate this cold-start latency for the predictable spike?
Provisioned concurrency pre-initializes a specified number of execution environments so invocations hit warm environments immediately, eliminating cold starts. Application Auto Scaling can scale provisioned concurrency on a schedule to match the known daily spike, which reserved concurrency alone does not do.
Question 4 of 6 · Development with AWS Services
A banking application must process withdrawal and deposit transactions for each customer account strictly in the order they occurred, but transactions for different customer accounts can be processed in any order relative to each other, and the system must handle thousands of accounts with high overall throughput. Which SQS configuration meets this requirement?
FIFO queues guarantee strict ordering within a message group. Using the account ID as MessageGroupId means each account's messages are delivered/processed in order among themselves, while separate groups (different accounts) can be processed in parallel, maximizing throughput.
Question 5 of 6 · Development with AWS Services
A developer is deciding between a DynamoDB Local Secondary Index (LSI) and a Global Secondary Index (GSI) for a new access pattern. Which statement correctly distinguishes the two?
LSIs share the table's partition key (only the sort key differs), must be defined at table creation, are limited to 5 per table, and uniquely among indexes support strongly consistent reads. GSIs can use any partition/sort key combination, can be added or removed anytime, but only support eventually consistent reads and have their own provisioned/on-demand capacity.
Question 6 of 6 · Development with AWS Services
A web application allows users to upload video files up to 2GB directly to S3 from the browser using presigned URLs. On unreliable mobile connections, uploads frequently fail partway through, forcing users to restart the entire upload from byte zero, causing high abandonment. Which change to the upload approach BEST solves this?
S3 multipart upload lets a large object be uploaded in independent parts (5MB minimum, except the last, up to 10,000 parts, 5TB object max), each with its own presigned URL. If a part fails, only that part needs to be retried — not the whole file — dramatically improving resilience on unreliable connections, and parts can even upload in parallel.
Ready for the real thing?
The full course: two full-length practice tests, video lessons for every exam domain, hands-on labs and detailed explanations.
$109.99$34.99 with code FREETEST33 — valid through September 19.