AWS Lambda

Amazon SNS now supports message payloads up to 1 MiB, up from 256 KiB, by setting the new MaximumMessageSize topic attribute on a topic. Topics with this attribute set above 256 KiB support AWS Lambda subscriptions (along with Amazon SQS and Amazon Data Firehose), allowing Lambda functions to receive larger SNS messages without prior payload splitting or offloading.

Matters for: Engineers building serverless applications that use Lambda functions as SNS subscribers to process application integration, IoT, or generative AI workloads with large message payloads.

AWS Lambda now lets you explicitly enable or disable direct read for S3 Files independent of function memory size, whereas previously direct reads from your S3 bucket were only available by default for functions with 512 MB of memory or higher. This is a non-breaking addition; no configuration changes are required unless you want to opt into or out of direct read behavior for specific throughput or latency needs.

Matters for: Teams building scalable data processing pipelines or stateful agentic workloads on Lambda that read files of varying sizes from S3.

Coding agents like Claude Code and Kiro can now inspect a running Lambda function and its connected resources across API Gateway, EventBridge, S3, DynamoDB, SNS, SQS, and Step Functions in a single call, correlating errors against a 7-day baseline, surfacing recurring error trends, retrieving deployed configurations, tracking a timeline of recent changes, and analyzing latency across those resources. This reduces token usage compared to orchestrating multiple separate API calls, and the capability is available at no additional cost through the Agent Toolkit for AWS or a standalone AWS MCP Server installation.

Matters for: Teams running production serverless architectures on AWS Lambda with multiple connected services who use AI coding agents such as Claude Code or Kiro for troubleshooting and operations.