A web application behind a CloudFront distribution needs near-real-time monitoring for error rates and anomaly detection. Which combination of actions should a developer take to build a dashboard that provides the most frequent monitoring? (Choose two.)

To achieve near-real-time monitoring, CloudFront real-time logs are essential as they deliver logs with minimal latency, making them suitable for immediate analysis. Integrating these logs with Amazon Kinesis Data Streams provides a highly scalable and durable data ingestion service capable of handling high-throughput log data. From Kinesis Data Streams, forwarding the data to Amazon OpenSearch Service via a Lambda function allows for flexible processing and transformation of the logs before indexing. OpenSearch Service is ideal for log analytics and anomaly detection, and its integrated OpenSearch Dashboards provide powerful visualization capabilities for building real-time dashboards. Streaming CloudFront logs to an S3 bucket and using Athena is suitable for historical analysis but not for near-real-time monitoring due to the batch nature of S3 log delivery and Athena queries. Kinesis Data Firehose is a good option for delivering logs to various destinations but doesn't offer the same real-time processing flexibility as Kinesis Data Streams combined with Lambda for custom transformations before OpenSearch. CloudTrail is for auditing AWS API calls, not for CloudFront access logs.