June 2019 Summaries

AWS API Gateway provides a powerful tool for creating a coherent API out of multiple disconnected remote function providers, but it introduces latency concerns that can be mitigated with proper configuration and alternative approaches. Common performance issues include cross-region function calls, resource-based versus role-based authentication, encryption/decryption of caches, and cold request starts, which can add hundreds of milliseconds to latency. Workarounds such as co-locating resources, modifying authentication, disabling cache encryption, and keeping serverless functions warm can help reduce the impact of these issues. Alternative approaches include hosting functions on EC2, containerizing serverless functions, evaluating other serverless hosts like Google Cloud Compute or Microsoft Azure Functions, and building a custom web server to bypass API Gateway translation. Ultimately, the choice between using AWS API Gateway with AWS Lambda depends on whether the latency introduced by this approach is too impactful to the user experience.

The author of the text has created a Lambda layer for an optimized version of the AWS SDK, which enables HTTP keep-alive and reduces latency. The layer is available at a specific ARN and can be included in projects using various frameworks such as Serverless or SAM. To manage the rollout and update of layers, the author has published an autodeploy-layer app that automatically adds the layer to functions in a region, allowing for easy deployment and maintenance. The app supports filtering by tags, enabling users to control which functions receive the specified layer. New versions of the optimized-aws-sdk layer will be published when the AWS SDK is updated, and can be rolled out to opted-in functions by redeploying the autodeploy-layer app with the updated ARN.

Lumigo, a serverless intelligence platform, has announced support for the Python microframework AWS Chalice, simplifying the process of writing serverless apps in Python. This addition follows recent launches of similar support for Java and Kotlin applications. Lumigo's integration with Osiris, a Kotlin framework, is also now possible, offering users more flexibility when building their serverless applications. As part of its ongoing development, Lumigo will continue to provide updates on new features, encouraging users to share their own requests through the provided contact information.

The author of the article is a traditional enterprise developer who has transitioned to serverless development and shares their journey and insights into this new architectural style. Serverless architecture offers opportunities for faster creation of products, scalability, security, and maintainability with less code. However, it also introduces new challenges such as debugging, configuration vs coding, cost estimation, testing, security, context switching, and the need for best practices. The author emphasizes that serverless is not a quick win but a journey that requires an open mind and preparation to adapt to new habits and learn new stuff. With its "Pay As You Go" pricing model, serverless can help save money in the long run but also raises concerns about cost estimation and potential security holes if not managed properly.

Erez Berkner from Lumigo is pleased to announce the official launch of their new Java tracer for AWS Lambda, extending their real-time monitoring and observability tools to JVM-based applications with support for multiple programming languages. The Lumigo platform continues to evolve with new features in development, including major announcements soon. Users are encouraged to provide feedback and suggestions through [email protected]. For more information on Lumigo's serverless application monitoring, a live demo can be requested today.