A team of data scientists needs a managed environment in which to prepare data, build, train, tune, and then deploy a custom fraud detection model to a hosted endpoint. Which AWS service is designed for this end-to-end workflow?
Intermediate · Choose one
Explanation
Answer: D. Amazon SageMaker AI (previously named Amazon SageMaker) is the AWS platform for the complete machine learning lifecycle. It provides managed notebooks, built-in algorithms, distributed training, automatic model tuning, and one-click deployment to managed inference endpoints, which is what building a custom model requires.
A. Amazon Athena is a SQL query service for data in Amazon S3. It can prepare data for machine learning but does not train or host models.
B. Amazon Rekognition is a pre-trained vision service. It cannot train an arbitrary fraud detection model on tabular transaction data.
C. Amazon Comprehend is a pre-trained natural language service. Its custom classification features are text-specific and do not provide a general model development platform.
32
A retailer keeps years of sales history in Amazon S3. It needs (1) a petabyte-scale warehouse that supports complex analytical SQL with high performance for repeated queries, and (2) interactive dashboards that business users can explore in a browser. Which TWO services should it use? (Choose TWO.)
Hard · Select 2 answers
Explanation
Answer: B and E. Amazon Redshift is a fully managed, petabyte-scale cloud data warehouse using columnar storage and massively parallel processing, which is what complex, repeated analytical SQL over large historical data sets calls for. Amazon Quick Sight (part of Amazon Quick Suite, formerly Amazon QuickSight) is the AWS business intelligence service, providing interactive dashboards and visualizations that business users access from a browser, and it connects natively to Redshift.
A. Amazon ElastiCache is an in-memory cache for reducing database read latency. It is not an analytics or reporting platform.
C. Amazon Kinesis Data Streams ingests real-time streaming data. The scenario describes analysis of years of stored history, not live ingestion.
D. Amazon DynamoDB is a NoSQL operational database optimized for high-volume key-value access, not for ad hoc analytical SQL across historical data.
33
A ride-sharing platform must ingest location events from millions of vehicles and have several independent applications process the same events within one second of arrival, with the ability to replay recent events if a consumer fails. Which AWS service best meets this requirement?
Hard · Choose one
Explanation
Answer: B. Amazon Kinesis Data Streams is built for real-time ingestion at massive scale. Records are retained in the stream for a configurable period, so multiple independent consumer applications can read the same records concurrently and a failed consumer can replay from an earlier position, all with sub-second latency.
A. Amazon Data Firehose (previously Amazon Kinesis Data Firehose) is a delivery service that buffers and loads streaming data into destinations such as Amazon S3 or Amazon Redshift in near real time. It does not retain a replayable stream for multiple independent consumers.
C. Amazon SQS is a message queue. In a standard queue each message is normally consumed and deleted by one consumer, so it does not naturally support several independent applications reading the same records or replaying them.
D. AWS Glue runs batch and streaming ETL jobs. It is a processing tool, not the ingestion and retention layer described here.
34
An operations team streams application logs from hundreds of servers into a central store. Engineers need to search across those logs interactively, build near-real-time operational dashboards, and raise alerts when error patterns spike. Which AWS analytics service is designed for this?
Hard · Choose one
Explanation
Answer: C. Amazon OpenSearch Service is a managed search and analytics engine built for exactly this pattern. It indexes log and event data as it arrives, supports fast full-text and structured search over that index, and includes OpenSearch Dashboards for near-real-time visualization and alerting, which makes it the standard AWS choice for centralized log analytics.
A. Amazon Redshift is a data warehouse optimized for complex analytical SQL over structured business data, typically loaded in batches. It is not a live log search and alerting platform.
B. Amazon Athena runs ad hoc SQL against data already at rest in Amazon S3. It suits periodic investigation rather than continuously indexed, near-real-time operational search and dashboards.
D. AWS Glue is a serverless ETL and data cataloging service. It can prepare log data for analysis but provides no interactive search interface, dashboards, or alerting.
35
Which AWS service provides a fully managed message queue that lets application components send, store, and receive messages so that they can operate independently and a failure in one component does not lose work?
Easy · Choose one
Explanation
Answer: A. Amazon SQS is a fully managed message queuing service used to decouple and scale microservices, distributed systems, and serverless applications. A producer places a message on the queue and a consumer retrieves it later, so the two components never need to be available at the same time and messages are retained until successfully processed.
B. Amazon SNS is a publish and subscribe messaging service that pushes messages to many subscribers at once. It is a fan-out mechanism rather than a durable work queue that a single worker pulls from.
C. Amazon EventBridge routes events from AWS services, SaaS applications, and custom sources to targets based on rules. It is an event bus rather than a pull-based work queue.
D. AWS Step Functions coordinates the order and state of multiple steps in a workflow. It orchestrates rather than buffering messages between components.
36
When a customer places an order, a company needs the same order event to be delivered simultaneously to an email notification system, an AWS Lambda function, and an Amazon SQS queue for the warehouse. Which AWS service provides this fan-out pattern?
Intermediate · Choose one
Explanation
Answer: C. Amazon SNS is a publish and subscribe service. A message published to an SNS topic is pushed to every subscriber of that topic, and supported subscriber types include email, AWS Lambda functions, Amazon SQS queues, HTTP endpoints, and mobile push. That one-to-many delivery is exactly the fan-out pattern described.
A. Amazon SES sends and receives email. It handles only the email destination, not delivery to Lambda and SQS.
B. Step Functions defines a sequenced workflow with explicit state transitions. It is not a broadcast messaging service.
D. Amazon SQS delivers each message to a single consumer that polls the queue. Achieving fan-out with SQS alone would require the producer to write to several queues itself.
37
A company wants to trigger automated actions when events occur in AWS services and in third-party SaaS applications, and it also needs to run some of those actions on a fixed schedule. Which service is designed as this serverless event bus?
Intermediate · Choose one
Explanation
Answer: A. Amazon EventBridge is a serverless event bus. It receives events from AWS services, custom applications, and integrated SaaS partners, matches them against rules, and routes them to targets such as Lambda functions, Step Functions state machines, and SQS queues. EventBridge Scheduler also supports time-based rules using cron and rate expressions.
B. AWS CloudTrail records API activity for governance and auditing. It can emit events that EventBridge acts on, but it does not route events to targets itself.
C. CloudWatch Logs stores and searches log data. Metric filters can raise alarms, but it is not a general event bus with SaaS integrations.
D. Amazon SQS stores messages for a consumer to poll. It has no rule-based routing, no SaaS event sources, and no scheduling capability.
38
A loan approval process consists of eight steps that must run in a defined order, with conditional branching, automatic retries on failure, and a visual record of exactly where each application is in the process. Which AWS service is best suited to coordinate this?
Hard · Choose one
Explanation
Answer: A. AWS Step Functions is a serverless workflow orchestration service. State machines define the sequence of steps, support choice states for branching, and provide built-in error handling, retry, and catch behavior. The console renders the workflow visually and shows the execution history for every run, which gives the required visibility.
B. A FIFO queue preserves message order but has no concept of workflow state, branching logic, or a visual execution history. The application would have to implement all of that itself.
C. AWS Batch schedules and runs batch compute jobs. It supports simple job dependencies but is not a general workflow engine with branching and visual tracking.
D. EventBridge rules route individual events to targets. They can chain reactions but do not maintain or visualize the state of a multi-step process.
39
A company must give 500 contractors access to a single specialized Windows engineering application. The contractors should not receive a full persistent desktop, and the application should be streamed to whatever device they already own. Which AWS service best fits?
Hard · Choose one
Explanation
Answer: D. Amazon AppStream 2.0 is a fully managed application streaming service. It delivers a specific desktop application to users through a web browser without giving them a full persistent desktop, and the application runs on AWS infrastructure while only pixels and input are sent to the device.
A. Running EC2 instances with Remote Desktop would require the company to build, license, scale, and secure the whole environment for 500 users manually.
B. Amazon WorkSpaces provides full persistent virtual desktops, which is more than the scenario requires and carries higher per-user cost and management effort.
C. Amazon WorkSpaces Secure Browser (formerly Amazon WorkSpaces Web) provides secure browser-only access to internal websites and SaaS applications. It does not stream an installed Windows desktop application.
40
A logistics company is launching two new capabilities: securely connecting and managing 100,000 delivery-vehicle sensors that publish telemetry to AWS over MQTT, and a cloud contact center that lets support agents handle inbound customer calls with automated routing. Which TWO services should it use? (Choose TWO.)
Hard · Select 2 answers
Explanation
Answer: B and E. AWS IoT Core is the managed service that lets connected devices interact securely with AWS and with each other at scale, handling device registration, authentication, and MQTT messaging for very large device fleets. Amazon Connect is the AWS cloud contact center service, providing inbound and outbound voice, interactive voice response, skills-based routing, and agent workspaces without any telephony infrastructure to run.
A. Amazon SES is a service for sending and receiving high-volume email. It handles a different communication channel and does not manage devices or route voice calls.
C. Amazon WorkSpaces provides managed virtual desktops for end users. It could give agents a desktop but provides none of the contact center or IoT capability described.
D. AWS Step Functions orchestrates the steps of a workflow. It could coordinate downstream processing but neither connects devices nor operates a contact center.