A media application uploads user photos to Amazon S3 for processing by AWS Lambda. Application state is stored in DynamoDB. When thousands of users upload at once, photo processing fails due to Lambda concurrency limits and DynamoDB write/read performance. Which combination of actions should increase performance and reliability? (Choose two.)

The problem states that DynamoDB write/read performance is an issue. Reviewing and adjusting Write Capacity Units (WCUs) directly addresses the write performance bottleneck, allowing DynamoDB to handle the increased load of user photo uploads. Similarly, if read performance were also an issue, adjusting Read Capacity Units (RCUs) would be necessary, but the primary bottleneck for new data is writes. Lambda concurrency limits are causing processing failures. Inserting an Amazon SQS queue between S3 and Lambda decouples the upload process from the processing. SQS queues can buffer messages (S3 event notifications), allowing Lambda to process them at its own pace, respecting concurrency limits. If Lambda fails to process a message, SQS can be configured to retry delivery, improving reliability. Adding an Amazon ElastiCache layer would primarily help with read-heavy workloads by caching frequently accessed data, but it doesn't directly address the Lambda concurrency or DynamoDB write performance issues. Enabling S3 Transfer Acceleration improves upload speed for users but doesn't resolve the downstream processing or database bottlenecks.