Question 310
A failed-event queue holds tax filings that can be manually repaired. The team wants to avoid discarding records yet also avoid endless retries of malformed payloads. What is the best disposition?
- Silently delete every message at its first transient failure.
- Use a DLQ with a bounded receive threshold, traceable investigation, and controlled redrive after repair.
- Retry malformed messages forever in the primary queue with no DLQ.
- Configure maxReceiveCount but leave the redrive destination unset and delete poison messages after repeated receive attempts.
Explanation
Correct Answer
B:Poison messages should be quarantined with evidence and replayed only once their cause is addressed.
Incorrect Answers
A:Transient failures can be retried safely and deletion discards valuable work.
C:Persistent errors consume worker capacity and may delay valid work.
D:A bounded count alone is not enough to preserve failed filings. A valid DLQ redrive policy must retain inspectable failed work and support deliberate replay after repair.
Question 311
Compliance must independently retain every payment-created event, while customer email can fail and retry without blocking compliance. Which event topology isolates both obligations?
- Send all events to one shared SQS queue with a consumer from each department.
- Publish once to SNS and provide a separate durable SQS subscription queue for each consumer.
- Invoke both downstream services synchronously and fail the payer if email is unavailable.
- Replicate an RDS table to the email service and assume that automatically guarantees an ordered event log.
Explanation
Correct Answer
B:Independent queues preserve fanout and each team’s retry/retention policy without sharing competing workers.
Incorrect Answers
A:Each message goes to a worker, not to every department.
C:Noncritical delivery availability should not block payment acceptance.
D:Database replication and event subscriptions have different contracts and replay semantics.
Question 312
An event bus must accept marketing and inventory changes, but one target should receive only inventory events. What boundary is appropriate?
- Give the inventory target a broad catch-all rule and ask the Lambda handler to discard marketing events.
- Keep both event classes on the bus and configure the consumer rule to match inventory source/detail fields.
- Use a separate EventBridge bus but assign a catch-all pattern to the inventory target so marketing events continue to invoke it.
- Use an SQS FIFO queue with only one group to guarantee business-category filtering.
Explanation
Correct Answer
B:EventBridge event patterns provide selective target routing independent of producer implementations.
Incorrect Answers
A:Downstream filtering wastes invokes and expands accidental processing risk.
C:Separating buses may help organization but the target’s rule must actually filter event source/detail to keep unrelated marketing events from arriving.
D:FIFO grouping controls order, not event-type eligibility.
Question 313
A once-weekly cleanup has no durable server that should stay online between runs, but the schedule must survive a maintenance deployment. Which hosting tradeoff is best?
- Use a cron task on a temporary development laptop.
- Call the cleanup API from every visitor’s HTTP request.
- Use an EC2 Auto Scaling lifecycle hook as the only calendar trigger, assuming a weekly maintenance task runs even when no instance changes lifecycle state.
- Use EventBridge Scheduler to invoke the approved cleanup target and persist its schedule configuration.
Explanation
Correct Answer
D:Managed scheduling removes dependence on a single instance process while retaining the specified calendar and retry behavior.
Incorrect Answers
A:The schedule disappears whenever the device is offline.
B:Unpredictable request arrival cannot guarantee a weekly execution.
C:Lifecycle hooks run on instance state transitions, not a durable wall-clock schedule. EventBridge Scheduler provides calendar-based invocation independent of an instance.
Question 314
A purchase workflow reserves inventory, authorizes payment and creates a shipping label. If payment authorization fails after an inventory reservation succeeds, the reservation must be released; if shipping-label generation fails, the workflow must record an approved recovery action. Which orchestration approach avoids an ad-hoc synchronous Lambda chain?
- Model explicit states, Catch/Retry, and separately authorized compensation using Step Functions.
- Reverse inventory reservations before knowing whether any inventory hold succeeded, and assume every later error is harmless.
- Launch validation, mutation and notification in parallel for minimum latency.
- Rely on every downstream API to always return HTTP 200.
Explanation
Correct Answer
A:A Step Functions workflow models validation, inventory reservation, payment authorization and shipping steps with explicit retries, Catch branches and narrowly scoped compensation for operations that actually committed. This makes failure handling and recovery visible.
Incorrect Answers
B:Compensating a business side effect before it is committed can invalidate legitimate work and does not solve a later payment or shipping failure. Compensation should be conditional on actual completed steps.
C:Mutation must not outrun authorization/validation.
D:Service failures and uncertain commit outcomes need explicit handling.
Question 315
A sensor emits tens of thousands of short independent telemetry transformations per second, each expected to finish within 20 seconds. No day-long execution history is required. Which Step Functions type merits consideration?
- Use Step Functions Standard only because Express cannot invoke any AWS Lambda functions, regardless of the sub-minute high-volume workload.
- Express with the claim it supports a multi-day wait for human approval.
- Express, with explicit downstream idempotency because asynchronous Express can execute steps at least once.
- Any workflow type without retries or duplicate guards because all service calls commit exactly once.
Explanation
Correct Answer
C:Express favors high event rates and short workflows; verify the exact Express invocation mode and duplicate semantics before accepting side effects.
Incorrect Answers
A:Both Standard and Express support appropriate Lambda integrations. The tradeoff is execution duration, volume, history and delivery semantics, not a blanket lack of integration support.
B:Express is duration-limited and inappropriate for days-long approval states.
D:Workflow execution guarantees do not automatically provide transactional exactly-once external effects.
Question 316
One Lambda function has two Amazon SQS event-source mappings: customer-orders and low-priority exports. The orders queue must never use more than 70 simultaneous invocations against the database, but exports should still have room to run. The function already has sufficient overall reserved concurrency. Which control isolates the orders queue contribution?
- Set MaximumConcurrency=70 on the orders SQS event-source mapping, keeping the function reserved concurrency high enough for all enabled sources.
- Set the function reserved concurrency to 70 and assume this separately limits just the orders mapping without impacting its exports mapping.
- Set the orders SQS queue visibility timeout to 70 seconds, assuming the number controls how many concurrent Lambda invocations its mapping generates.
- Raise the function provisioned concurrency to 1,000 and assume preinitialized environments cap the order source at 70.
Explanation
Correct Answer
A:SQS event-source scaling configuration caps concurrent invokes contributed by that specific mapping, whereas the function-wide reserved ceiling remains large enough for the combined event sources.
Incorrect Answers
B:Function reserved concurrency constrains all of the function’s invocation sources combined. It cannot assign a 70-invocation cap only to one queue while guaranteeing room for unrelated exports.
C:Visibility duration governs message redelivery protection, not the mapping’s simultaneous invocation limit. Configure the SQS event-source MaximumConcurrency for the queue.
D:Provisioned concurrency preinitializes execution environments to lower latency; it does not establish a per-queue max concurrent invocation ceiling.
Question 317
An SQS FIFO queue invokes Lambda with a ten-message batch from one message group. The handler uses ReportBatchItemFailures and encounters an invalid record at position four. How should it report the batch to preserve ordered processing?
- Continue processing records five through ten in the same group and mark only record four failed, even though later records depend on its state.
- Throw an unhandled exception after processing the first three records and expect Lambda to acknowledge only the first three by default.
- Stop processing at record four and return the failed record plus all remaining unprocessed message identifiers in batchItemFailures.
- Set BatchSize to 10,000 so the ordering guarantee automatically ignores failed FIFO messages.
Explanation
Correct Answer
C:AWS recommends stopping after the first FIFO failure and reporting failed and unprocessed message IDs in the valid partial batch response so earlier successes can be acknowledged and ordered replay preserved.
Incorrect Answers
A:For FIFO ordering, records after the first failure in a message group should not be processed ahead of that failure. Marking only the failed record can permit unsafe advancement of later dependent messages.
B:An unhandled Lambda exception makes the entire batch a failure; default event-source mapping behavior can redeliver records previously handled successfully.
D:SQS FIFO event-source batches have their own batch-size limits, and a larger batch would not implement the necessary failure reporting or FIFO ordering safety.
Question 318
A mobile API needs JWT-aware authorization, per-client throttling and a serverless Lambda backend, but no long-running application reverse proxy. Which edge API component best matches the requirement?
- Expose Lambda through a public S3 bucket policy to authenticate API callers.
- Expose supported API Gateway methods or routes with correct authorizers, throttling and Lambda integration.
- Deploy one always-on EC2 web server solely to duplicate managed API Gateway primitives.
- Use Route 53 weighted latency records plus a Cognito user pool but no HTTP API authorizer integration, expecting DNS to validate JWT claims per request.
Explanation
Correct Answer
B:API Gateway is a managed HTTP API front door; ensure feature support in the selected REST/HTTP API type.
Incorrect Answers
A:S3 policies do not serve as JWT-aware API routing and authorizer controls.
C:An EC2 proxy is possible but adds ops and undermines the specified serverless design.
D:DNS routing does not inspect JWTs or enforce per-client API throttles. API Gateway must be configured with its supported authorizer and integration semantics.
Question 319
An ALB listener has a rule requiring BOTH Host=pay.example.com and Path=/checkout/*. Requests for https://shop.example.com/checkout/cart unexpectedly follow the default target group, while https://pay.example.com/checkout/cart reaches checkout. Which edit fixes the domain-specific mismatch?
- Increase the checkout target-group health-check timeout; healthy targets make ALB host-header conditions optional.
- Add shop.example.com to the intended host-header match or adjust the rule design, while retaining the /checkout/* path condition and verifying the expected rule priority.
- Replace the ALB with a Network Load Balancer because ALBs cannot combine host-header and path conditions.
- Change a Route 53 weighted DNS record but leave the host-header condition unmatched, assuming DNS decides the listener rule after the request reaches the ALB.
Explanation
Correct Answer
B:ALB listener conditions are conjunctive across condition types. Matching /checkout/* alone is insufficient when the Host header is shop.example.com but the rule permits only pay.example.com.
Incorrect Answers
A:Target health affects forwarding after a listener rule matches; it cannot make the submitted Host value satisfy a different hostname condition.
C:ALBs support host-header and path-based routing; an NLB is layer 4 and does not solve the application listener condition mismatch.
D:DNS selects an endpoint for a name; the HTTP Host header and listener conditions still determine which ALB rule applies after connection.
Question 320
An ECS service runs successfully on Fargate, but new tasks placed in private subnets fail to pull a private image from Amazon ECR. The task execution role is already correctly authorized, and the subnets have no NAT or Internet egress. Which connectivity design should the team check?
- Provide the relevant ECR API and Docker interface VPC endpoints and S3 gateway endpoint with usable private DNS, route-table and security-group settings, or provide authorized NAT egress.
- Move from Fargate to EC2 launch type because ECS tasks cannot pull ECR images when placed in private subnets.
- Assign a larger CPU value to the Fargate task definition because insufficient CPU makes ECR endpoints public-only.
- Create a new IAM Identity Center permission set for end users and assume workforce SSO automatically supplies Fargate image-pull network routing.
Explanation
Correct Answer
A:Private Fargate tasks still need a network path to their registry and image layers. Supported ECR interface endpoints plus S3 gateway routing (and any required logging/secrets endpoints) or NAT can supply connectivity without public task IPs.
Incorrect Answers
B:ECS on Fargate supports private tasks and ECR image pulls, provided execution-role permission and service access paths are properly configured. EC2 host management is not required.
C:Increasing task compute capacity cannot create missing private network routes, VPC endpoints or Internet translation needed to fetch container image layers.
D:Workforce identity provisioning does not establish the task ENI routing or service DNS paths required by ECR image pull requests.
AWS Certified Solutions Architect - Associate SAA-C03 Practice Tests