Off-by-none: Issue #206

October 25, 2022

Only have a few minutes? Check out this week's MOST POPULAR links as chosen by our email subscribers.

You're doing cloud wrong 🙅

Welcome to Issue #206 of Off-by-none (Community Edition)! 🎉 This issue is sponsored by our friends at Momento.

In our previous issue, Webiny launched a Headless CMS+, AWS upgraded their event source mappings, and we talked about serverless platform engineering. This week, Cloudflare brings back server-side includes, Oracle launches “serverless Kubernetes”, and we look at why some people are just plain doing cloud wrong. Plus, we have plenty of great content from the serverless community.

Alex DeBrie breaks down 3 choices that determine the perfect caching strategy
Caching is fast (hello, sub-millisecond p99 response times). Caching is also fun (nothing is more fun than delighted users and repeat customers). But caching is tricky, meaning the right caching strategy is key. Determining strategy starts with 3 simple questions: where, when, and how? Read Alex’s full thoughts on the Momento blog. SPONSORED

Serverless News & Announcements 📣

Cloudflare Workers launched a new fragment architecture to support micro-frontends, or as Michael Hart cleverly put it, “server-side includes.” In addition to that (and I’m sure several other things Cloudflare is doing), the Next.js Edge Runtime became the fourth full-stack framework supported by Cloudflare Pages.

In more “other” cloud news, Oracle introduced support for “Serverless Kubernetes” on Oracle Cloud Infrastructure. 🤦‍♂️ Somebody please stop the ride, I want to get off! More sensically, Azure added sustainability guidance to their Well-Architected Framework.

In AWS news, they announced dark mode support in the AWS Management Console. This was surprisingly (or maybe not surprisingly 🤔) the highlight of many peoples’ week.

And then there was the announcement for the AWS Parameters and Secrets Lambda Extension. I hate to criticize, because I like the idea of implementing a best practice like this, but the interface for it is pretty terrible. A local http server makes sense so it can be supported across runtimes, but making an http call seems messy. Would have liked to see runtime specific implementations of this that didn’t require passing in session tokens and storing configs in environment variables.

If you want to see how it works, take a look at How to use AWS Parameters and Secrets Lambda Extension and Antonio Lagrotteria’s AWS Parameters and Secrets Lambda Extensions Usage.

Serverless Concepts 🏗

Kieran Wrenn has an excellent post on why you should use a single API Gateway endpoint per microservice. I wholeheartedly agree with this. If you want a unified endpoint for multiple microservices, put regional APIs behind a single CloudFront distribution.

Jones Zachariah Noel explains why you should use AWS Step Functions and SDK integrations.

Luca Mezzalira and friends share some advice when architecting for NoSQL databases.

If you want to geek out a bit on DynamoDB, this Twitter thread from Akshat Vig explains how they built tables that adapt to changing traffic patterns.

And finally, Justin Perkins tackles Lambda defense in depth with a few key cloud security concepts to elevate your Lambda function security.

Serverless Tutorials 👷‍♀️

Lots of interesting tutorials as always. Here are a few that caught my attention.

Hannes Borch has a great tutorial on optimizing AWS Lambda for mass-data import. The best advice I’ve heard is that Lambda is best for transforming data, not for transporting it. Sometimes Lambda functions are necessary, but if there is a service that can accomplish what you need, take a look at that.

Matt Martz shows how to use the AWS CDK to deploy EventCatalog, an open source project built by David Boyne that helps you document your events, services and domains. Matt also has another great post that shows you how to protect a static site with Auth0 using Lambda@Edge and CloudFront.

Maybe not the most serverlessy thing in today’s newsletter, but Dave Hall demonstrates how he tracks infrastructure with SSM and Terraform. The concept could be used with other frameworks and deployment tools as well.

Bit of a mind bender here from Max Rohde with his Serverless React SSR post. He walks you through developing a truly serverless implementation for React SSR and highlights the good, bad, and ugly.

JV Roig shared part 3 of his Serverless Analytics series. This one discusses how to achieve 90% faster and 80% cheaper queries through cataloging and crawling.

Maurice Borgmeier explains how to implement pessimistic locking with DynamoDB and Python, and this other post from Aniketh Deshpande explains both optimistic and pessimistic locking. Both methods have their uses, but I tend to prefer optimistic locking when the likelihood of collisions is low.

Finally, Sai Rahul shows you how to perform realtime transformations on S3 with Amazon S3 Object Lambdas.

Serverless Reads 🤓

Mattias Andersson’s take on why people need to rightly move away from the ‘lift and shift’ approach to cloud migration.

Erik Bernhardsson from Modal has a great piece explaining that we are still early with the cloud and software development is overdue for a change. Couldn’t agree more.

I had some similar thoughts in my Cloud Native versus Native Cloud apps post.

Allen Helton shares an insightful post about the power of high functioning dev teams.

This piece on Cloud-native strategies for digital business acceleration highlights, perhaps inadvertently, the dichotomy of CIOs and CTOs “realizing the importance of containerized development” but also the “need to leverage their cloud providers’ specific services and tools.” The former is the past, and is exactly what is holding us back from realizing the true power of the cloud.

Podcasts, Videos, and more 🎧

There was a great talk at the Strange Loop Conference by Jack Rusher called “Stop Writing Dead Programs”. It’s not about serverless, but there are some really interesting thoughts in there about the importance and power of dynamic runtime feedback versus compiled code. Just because we’ve been doing it one way for so long, doesn’t mean there isn’t a better path forward.

New from AWS 🆕

Lots of announcements from AWS this week as they slowly open the pre:Invent firehose spigot.

• Amazon CloudFront adds fields for origin latency and ASN in real-time logs for more granular insights
• AWS Step Functions adds new execution observability features for Express Workflows
• Visualize application health with Amazon CloudWatch Application Insights
• Amazon Cognito now provides user pool deletion protection
• Amazon Aurora MySQL supports faster export to S3
• Amazon SQS announces increased throughput quota for FIFO High Throughput (HT) mode to up to 6,000 Transactions Per Second (TPS)
• AWS Global Accelerator announces AddEndpoint and RemoveEndpoint APIs
• IAM Identity Center adds session management features for improved user experience and cloud security
• Amazon SES now offers new model to simplify provisioning and managing dedicated IPs
• AWS Batch increases compute and memory resource configurations for Fargate type jobs by 4X
• Announcing updates to the AWS Well-Architected Framework
• Amazon S3 Replication now supports SSE-C encrypted objects
• Announcing new AWS Amplify Library for Swift, now with support for both iOS and macOS

Thoughts from Twitter 🐦

I spoke with a team yesterday that is moving away from serverless in favor of managed k8s on aws

it wasn't an easy decision for them and they went through a thorough review of their architecture

the reasons they highlighted to me are:

— Boris Tane 🍋 (@BorisTane) October 21, 2022

Great thread from Boris that outlines a company’s reason for abandoning their serverless approach. The comments and the quoted tweets are also filled with lots of insight worth exploring. And be sure not to miss Yan Cui’s thoughts on this as well.

The cloud isn’t the issue, you’re just using it wrong…

Last week, tech Twitter was all abuzz after DHH boldly proclaimed and explained why [37signals is] leaving the cloud. A lot of people cheered, some of us jeered, and everyone else just pitched web3 as an alternative solution. DHH’s success has earned him a giant platform and a tremendous amount of influence, and while I often disagree with him, it’s clear that many others do not. I spent quite a bit of time reading through all the retweets, reposts, comments, and hot takes, and I came to a fairly simple conclusion: these people are using the cloud wrong.

I shared more thoughts in this week's premium edition. 

Upcoming Serverless Events 🗓

If you have an event, webinar, etc. that you’d like me to mention, please email me.

October 27, 2022 – Building Multi-tenant SaaS Applications on AWS

November 28 – December 2, 2022 – AWS re:Invent

Serverless Star of the Week ⭐️

There is a very long list of people who are doing #ServerlessGood and contributing to the Serverless community. These people deserve recognition for their efforts. So each week, I will mention someone whose recent contribution really stood out to me. I love meeting new people, so if you know someone who deserves recognition, please nominate them.

This week’s star is Michele Sciabarrà (@sciabarracom). Michele is the founder and CEO at Nuvolaris.io, the Serverless Anywhere company. Previously, he was Regional Manager Europe for Nimbella (acquired by DigitalOcean) and had founded a consultancy focused on Kubernetes and Serverless solutions. Michele is the author of Learning Apache OpenWhisk: Developing Open Serverless Solutions, and Linux e Programmazione Web (for those Italian speakers in the community!) He’s also the Apache Software Foundation PMC of the OpenWhisk Project. Thanks, Michele, for your contributions to the serverless discourse!

Final Thoughts 🤔

I hope you’re enjoying the community edition of the newsletter! If you want to support the newsletter and get extended commentary every week, be sure to check out the premium edition.

Take care,
Jeremy

I hope you enjoyed this newsletter. We’re always looking for ideas and feedback to make it better and more inclusive, so please feel free to reach out to me via Twitter, LinkedIn, or email.