#​202 — May 28, 2021

Read on the Web

Serverless Status
Serverless news, views, and developments each Friday

The State of Serverless 2021 — In early 2020 we linked to a report from Datadog called State of Serverless looking at trends in the serverless space collected together from their thousands of users. It’s back in a 2021 form with a variety of added insights:

  • Azure Functions and Google Cloud Functions haven't exactly caught up with AWS Lambda but are gaining momentum with a larger share of companies using them than before.
  • Lambda function use has grown, unsurprisingly, with functions now being invoked 3.5x more than two years ago.
  • Pay as you go pricing clearly works as Lambda invocations are now far shorter than even a year ago at about 60 milliseconds on (median) average.
  • Serverless Framework continues to dominate in the tooling and deployment space.
  • In percentage share, Python use has grown, JavaScript has dipped.


Application Monitoring Built for Serverless — See performance bottlenecks and exceptions in distributed applications by analyzing request flows with payloads, metrics and events. Search any payload, users, or tags in application-level calls; explore end-to-end request flows across microservices.

Epsagon sponsor

AWS Lambda Extensions Now Generally Available — AWS Lambda Extensions provide a way to integrate Lambda with external observability and monitoring tools. A handful of providers offered extensions during the preview phase but it can now be used more broadly, including for sending telemetry to custom destinations after a function has finished running.

Channy Yun (AWS)

Building a Serverless Outbreak Management Solution with AWS Data Lab — Unsurprisingly I’ve never needed to do this, but a real life case study is always appreciated as a learning exercise. Building a portal with React and AWS Amplify helped a manager of retirement communities keep a handle on COVID metrics.

Mondal, Mittal and De Jesus

Five Easy Steps to Deploy Containers as Lambda Functions — A practical introduction to using containers as the vector for getting things running on AWS Lambda.

Varun Singh

Convect: Instant Serverless Deployment of Machine Learning Models — A new platform for deploying ML models from Jupyter Notebooks. A good example of taking the serverless concept to an audience who simply isn’t paid to care about ops.

Convect Inc.

Behind the Scenes as AWS AppConfig Builds a Lambda Extension — Steve Rice and Tim Yao share “why the AWS AppConfig team built an AWS Lambda extension, the effort required to build it, and the outcomes”.

Amazon Web Services

How to Create and Publish OpenAPI-enabled Azure Functions with Visual Studio and .NET
Shayne Boyer

Stateful Serverless Automation with PowerShell Support in Azure Durable Functions
Anthony Chu (Microsoft)

Seen on Twitter..

Spencer's question provoked a lot of interesting responses covering the gamut from things like Fauna and DynamoDB to using a managed traditional database (e.g. Postgres) or Azure Cosmos DB.