Our serverless superhero this week is David Boyne, an AWS Community Builder who also happens to be an EventBridge whisperer. He manages several open source projects and regularly contributes to the serverless community. Seriously, if you need help with event-driven architectures, David is your guy. Join me in giving him a big thank you for what you do for the community.
AWS has provided us with not one, but two major service updates this week.
The first is the introduction of Lambda function URLs, which enables users to get either a public or IAM secured endpoint to directly invoke the function. This eliminates the need for API Gateway in certain use cases and opens the door for some enhanced service to service communications in a loosely coupled manner.
The other is the introduction of EventBridge global endpoints, which provides automatic region failover for event processing in case of an outage. This is huge for companies running global applications who are handling this manually already.
In part 5 of his Lambda best practices series, Ran Isenberg talks about how to do input validation for your Lambda functions. This is particularly useful because of the many different inbound schemas from the array of event triggers in AWS.
With the introduction of Lambda function URLs, a slew of blog posts came with it. AJ Stuyvenberg wrote up a thorough, in-depth analysis of the new functionality, detailing what it does and does not do. If you’re looking to understand what the function URLs stand to provide, take a look at his article.
Wrapping up my spotlight on Step Functions, I (Allen Helton) conducted a series of tests to compare Lambda and Step Functions cost and performance. This post walks through costs for synchronous and asynchronous workflows written in Lambdas and as state machines. This article finally helps solve the which is better, Lambda or Step Functions question.
Zoran Ivanovic has a thought provoking article on using parallelism to optimize large queries in DynamoDB. He explains how to break down queries, provides code samples, and runs through a thorough amount of test cases to prove his point.
As part of the AWS Community Day in Turkey last week, Jeremy Daly gave a presentation on DynamoDB modeling for serverless applications where he covers everything from AWS concepts to how to think differently with NoSQL vs SQL to identifying and maintaining access patterns.
It’s easy to rush into serverless because of its low barrier to entry. And oftentimes, we see companies dive in without the necessary governance in place to make it successful. Matt Coulter talks about the mentality shift that needs to be made through all parts of the organization. In his writeup, he talks about enabling architectures, safety nets, and using AWS tools to help you succeed.
From the recent trends we’re seeing from AWS and the serverless community, it’s clear to see there are two things on everyone’s minds: availability and performance optimization. When you think about it, those are two of the most important components of serverless. Put some effort in implementing some of the phenomenal work listed above. It’s where the serverless community is heading.