A company’s IoT platform currently runs on-premises: a server uses MQTT to collect telemetry from devices every 5 minutes and stores device metadata in a MongoDB cluster. A periodically running on-premises app aggregates and transforms telemetry+metadata to produce reports; those jobs run 120–600 seconds. A separate web app always runs to serve reports. The company is moving to AWS and wants to reduce operational overhead. Which combination of steps will meet the requirements with the LEAST operational overhead? (Choose three.)

Configuring IoT devices to publish to AWS IoT Core directly leverages a fully managed service for ingesting device telemetry, eliminating the operational overhead of maintaining on-premises MQTT brokers. Writing device metadata to Amazon DocumentDB (with MongoDB compatibility) replaces a self-managed MongoDB cluster with a fully managed, scalable, and highly available database service, significantly reducing operational burden compared to running MongoDB on EC2. Using AWS Step Functions with Lambda tasks for report generation and storing reports in Amazon S3, served via CloudFront, creates a serverless, scalable, and cost-effective reporting pipeline. This eliminates the need to manage servers for report processing and serving. Replacing on-premises connectors with Lambda functions to connect directly to IoT devices is generally not feasible or recommended; devices typically publish data, not receive connections from a serverless function. Using a self-managed MongoDB on EC2 increases operational overhead compared to DocumentDB. An Amazon EKS cluster introduces significant operational complexity for managing Kubernetes, which is unnecessary for the described reporting workload and serving static reports.