May 2021 Summaries

AWS Lambda Extensions are now Generally Available in US East (N. Virginia), EU (Ireland), and EU (Milano) regions, with more regions to come, and offer operational tools to integrate with Lambda functions, allowing them to run either in-process or in a separate process, with improved performance and the ability for extensions to return early from function invocations, reducing latency and enabling real-time data processing. However, authors must consider that they will still pay for compute time used by extensions after the function response is returned to the caller, and slow extensions can affect cold start frequency.

The text discusses ways to improve the performance of AWS Lambda functions, specifically focusing on cold starts and timeouts. Cold starts occur when a new Lambda worker is created, taking longer to process requests due to initialization time. To mitigate this, it's recommended to use Provisioned Concurrency or optimize function code to reduce module initialization time. Debugging Lambda timeouts involves identifying the root cause, which can be elevated latency from external services like DynamoDB. To address this, it's essential to track latency metrics for dependencies and use tools like Lumigo or X-Ray to automate tracing and debugging. These tools provide a more polished developer experience and support async event sources, TCP-based transactions, and logging capabilities that make it easier to identify and debug issues in complex serverless applications.