Free AWS Certified Solutions Architect - Associate practice questions

Question 01

A bank wants to deploy a stateless containerized notification API directly from a container image with automatic scaling, load balancing, and HTTPS, without provisioning or managing clusters, load balancers, or ECS task definitions. Which service best meets this goal?

Question 02

A bank's online-banking web tier runs on EC2 behind an ALB in an Auto Scaling group. Occasionally the JVM deadlocks: the instance still passes EC2 status checks and shows as running, but it returns HTTP 503 and stops serving customers. The bank wants the Auto Scaling group to automatically replace instances in this stuck state. What should be configured?

Question 03

A bank wants a managed way to be automatically alerted when spending unexpectedly deviates from normal patterns for any service, and to have a budget automatically apply a corrective action when a threshold is breached. Which combination should be implemented?

Question 04

A company runs several always-on Lambda functions that together account for a large share of its monthly compute bill. The functions are pure compute in Node.js, contain no native binaries tied to a specific processor architecture, and can be rebuilt easily. The company wants to lower Lambda cost with minimal effort and no architectural redesign. What should the solutions architect recommend?

Question 05

A bank runs a background media-transcoding service on Amazon ECS where each task is stateless, writes output to S3, can be safely restarted, has a deep queue, and relaxed deadlines. It wants to minimize the compute cost of running these tasks on serverless containers. Which launch configuration should it use?

Question 06

A bank must ensure that any S3 bucket in an account that becomes publicly accessible is automatically remediated within minutes by re-enabling S3 Block Public Access, with continuous detection and no human in the loop. Which combination meets this requirement?

Question 07

A bank runs a customer-facing API on AWS Lambda behind API Gateway and must protect a fragile downstream mainframe adapter that can handle only a bounded number of concurrent calls, while also ensuring no cold-start spikes push a subset of requests past a 250 ms SLA during brief surges. Which configuration meets both needs?

Question 08

An application writes orders to an Amazon DynamoDB table. A separate team needs a continuously updated, materialized aggregate view (for example, per-region order totals) derived from every item change, updated within seconds and without the ingestion application polling the table or running scheduled scans. Which approach meets these requirements?

Question 09

A bank must ensure that an S3 bucket used by a machine learning pipeline can be read only by principals in the bank's own AWS Organization, blocking any principal outside the organization even if a resource policy is misconfigured elsewhere. Which condition enforces this on the bucket?

Question 10

A bank needs disaster recovery for a critical web application into a second Region with an RTO of about 10 minutes and an RPO of 1 minute, but wants to avoid running a full-scale duplicate of production continuously. The DR environment may run at reduced capacity and scale up on failover. Which DR strategy best fits?

Question 11

A company runs a bursty CI/CD build system on Amazon EKS. The build pods start and stop constantly, pod density varies widely through the day, and there are long idle stretches overnight with no builds at all. The platform team does not want to pay for or manage worker nodes that sit partly empty during quiet periods and wants to pay only for the resources each pod actually requests while it runs. Which EKS compute option best minimizes cost for this pattern?

Question 12

A bank serves a global single-page banking app whose static assets live in two S3 buckets, us-east-1 (primary) and eu-west-1 (secondary), kept in sync by replication and delivered through Amazon CloudFront. The bank needs CloudFront to automatically fetch from the secondary bucket if the primary returns 5xx errors or times out. Which configuration meets this?