The serverless takeover…
Welcome to Issue #21 of Off-by-none. I hope you’re ready to talk serverless! 😃
Last week we got hands-on and learned how to handle “not-so-scalable” systems in our serverless applications. This week we look at some more ways to scale your serverless apps, highlight some recent innovations, examine how serverless and the cloud is affecting the IT landscape, and so much more.
Lots to get to, so let’s jump right in! 🏊♂️
When you’re trying to get your serverless application to scale… 📈
Mikhail Shilkov has a brilliant post titled: Serverless at Scale: Serving StackOverflow-like Traffic. In this post he runs experiments across AWS, GCP, and Azure, to test how serverless functions and blob-storage scales to 1,000 requests per second. The results are quite fascinating.
We often talk about scaling “non-serverless” downstream systems in this newsletter, and Tirumarai Selvan has presented us with another option for Scaling RDBMS for GraphQL backends on serverless. Connection management is an ongoing problem with serverless functions. AWS is working to fix this with their Data API for Aurora Serverless (and of course there’s my serverless-mysql package), but overall, not a bad (albeit, non-serverless) approach.
Paul Johnston has some thoughts on Serverless Compute and Serverless Data. It is an interesting way to compartmentalize serverless applications. Without the proper design, ephemeral compute is certainly limited by the underlying datastore. Designing for scale is the new default, and this is a skill that many developers have never really needed to worry about.
Tim Bray started this thread on Twitter that goes deep into microservices and temporal coupling through synchronous communication. 🤓 I love these types of discussions, especially when Marc Brooker and Sam Newman jump in.
And James Thomas tells us about loosely-coupled serverless functions with Apache Openwhisk. Good read that looks at the difference between triggers and queues and how they can affect the scalability of your severless application. A bit specific to Openwhisk, but I think the general concepts are quite universal.
When people are having way too much fun with custom runtimes… 👩💻
aws-sdk with your Lambda functions. This way is much better. 👍
The team over at Thundra developed their own Node.js Custom Runtime to let you monitor your Lambda functions without making any changes to your code. We’ve seen this type of use case before, but Thundra went the extra step to show us how they actually built it.
Have you ever wondered how to run Elixir on Lambda? Me neither, but Arjan Molenaar has figured it out for us just in case. Building an Elixir runtime for AWS Lambda gives you a brief overview of his motivations, and ultimately leads you to the GitHub repository if you’d like to try it yourself.
And PHP fans can also rejoice! Bref, a serverless framework for PHP, is incorporating a custom PHP runtime into v0.3. Look forward to better performance, PHP-FPM support, and local development with Docker and AWS SAM.
Where to look for serverless events… 🗓
ServerlessDays Cardiff is coming up on January 30th. Tickets are still available, so if you’re going to be in the area, I’d highly suggest you attend. Can’t go wrong with talks from the likes of Yan Cui, Simona Cotin, the Ian Massingham, Slobodan Stojanović and so many more.
And if you’re state-side, ServerlessDays Boston is coming up on March 12th. We just announced the one and only Charity Majors as our opening keynote speaker. And I’m happy to announce that the, wait for it… legendary Chris Munns from AWS will be giving the closing keynote. The remaining speakers will be announced early next week. This is going to be good. 🙌
If you’re looking for something a bit more remote-friendly, Stackery has some upcoming serverless webinars that you can join. They’ll walk you through how to build your serverless applications without needing to write a bunch of YAML.
Feel like doing some traveling? Thundra put together a great list of Serverless Events You Should Be Aware Of in 2019. I’m going to try and get to a few of these myself.
For those of you that are visual learners… 👀
I stumbled across some videos that Cloud Path had created, and I was impressed with how well-produced they were. In AWS S3 & AWS Lambda Integration, they walk you through setting up an S3 trigger and the code required to process the event. Beginner level stuff, but I’m going to keep my eye on this channel.
Marcia Villalba dropped another re:Invent interview where she’s Talking about testing Serverless applications with Slobodan Stojonovic. Slobodan was our very first Serverless Star at Off-by-none and is an awesome serverless resource.
If you can’t get enough of Marcia, check out her Getting ready for AWS reInvent 2018 vlog series. If you’re thinking about going to re:Invent this year, these videos provide a first hand look at this amazing experience.
CloudFlare workers are a relatively new addition to the serverless ecosystem, and they’re quite passionate about how this type of edge computing could change how applications run. How Serverless Platforms are Changing to Enable New Applications is a talk by Zack Bloom that digs deep into this concept.
What to do if you’ve been ignoring serverless security and user privacy… 🔒
If you thought that you didn’t need to worry about GDPR, think again. It was just reported that France fined Google nearly $57 million for an alleged violation. Now this might just be France being France, or it’s a sign of things to come. If you’re not familiar with GDPR, or you’ve already forgot the requirements, Stripe has a great guide to help you out. C’est la vie. 🇫🇷
Last time I’ll mention this (promise). Ory Segal and I are hosting a Foundations of Lambda Security webinar on January 24, 2019 at 11am ET. It will be packed full of practical serverless security advice including risks associate with AWS Lambda, IAM permissions, governance and regulatory compliance, and scalability.
When you’re looking for innovation in the serverless ecosystem… 🔍
Epsagon continues to make serverless observability easier with the introduction of Trace Search. This is a very cool feature that lets you find and drill down into traces using a bunch of different filters. Plus they have created plug-in packages to make integrating tracing and cleaning up your old Lambda versions much easier.
But serverless observability and tracing is a hot space to be in, and Adam Johnson and the team over at IOpipe has their own long list of accomplishments and future plans. In Auld Lang Servers, Adam outlines IOpipe’s milestones and innovations over the last year. Their product continues to get better and better, giving serverless practitioners plenty of options when choosing an observability tool.
And don’t count out OpenWhisk. Release 0.17.0 (18.01.2019) of the Serverless Framework OpenWhisk plugin was recently released, with added support for concurrent actions, which should speed up your deployments.
When you find out that Google Cloud Functions finally supports Go… 🤷♂️
Google announced that Go 1.11 is now a supported language for Google Cloud Functions. You’d think that since they invented it, they might have beat Amazon to the punch. Oh well, at least GCP is still innovating its serverless offerings.
Not to be outdone by AWS’s classic serverless example, Adil H has put together a post showing us how to do Image Resizing with Go and Cloud Functions. Code included.
If you’re looking to push the envelope a bit more, Saurabh Deoras has a great article on combining TensorFlow, Go and Cloud Functions. I like when people experiment with stuff like this, and even though his final solution isn’t ideal, it still works. He even waxes-poetic at the end. #deep
When the zombie apocalypse might not be the apocalypse you need to worry about… 🧟♂️
Forrest Brazeal wrote a rather depressing (but necessary) piece about the The Creeping IT Apocalypse. With AWS reportedly working on a secretive low-code/no-code project, there is an entire class of engineers that could get automated out of existence. TLDR; learn to code and keep your skills current.
Along the same lines, James Beswick’s latest post, The cloud skills shortage and the unemployed army of the certified, comes at it from a slightly different angle. Of course IT head counts are dropping because of automation, but James argues it isn’t just about keeping your skills current. It’s about the unreasonable expectation that a single developer must now do the jobs of what used to require several highly-specialized people to do. TLDR; become a coding superstar.
Other people are writing about this trend, perhaps without even realizing it. Nader Dabit gives his take on what it means to do Full-Stack Development in the Era of Serverless Computing. “This means you basically have a team of specialized engineers that have built out and iterated on something that you or your team simply could not do alone without investing an impractical number of hours.” I think this type of innovation is great, but don’t get caught watching shadows on the wall, this type of undifferentiated development work is going away. Now look who’s being poetic. 😉
When you really like seeing serverless use cases… 🤗
I think we are all in agreement that CloudWatch is not the best place to be digging into our application logs. There are plenty of options out there, but the team at BBC iPlayer shows us how they put Lambda Logs in ELK. It’s a DIY option, but highly effective for their needs.
This is a bit of an old post, but in How I export, analyze, and resurface my Kindle highlights, Sawyer Hollenshead show us how he created a serverless pipeline that extracted his highlights, analyzed them with NLP, and published them to his site to reflect on what he read. Pretty interesting use case, IMO.
Gavin Lewis shows us How To Build a Serverless CI/CD Pipeline On AWS. There is quite a bit of complexity to his approach, but he has it all laid out for you.
When you’re a big fan of the horror genre… 👹
Henning Jacobs has compiled a list of wonderful Kubernetes Failure Stories for us. He claims that these stories “should make it easier for people dealing with Kubernetes operations… to learn from others and reduce the unknown unknowns of running Kubernetes in production.” I say it’s just another opportunity for serverless fans to say I told you so 😂. But seriously, if you want to take a stab at Kubernetes, this is a good list to get you started (or maybe scare you away).
Corey Quinn recounts a horror story of his own in this Twitter thread. The story of an ambitious young man trying to set up his own infrastructure in a shared datacenter goes horribly awry, hilarity ensues. I remember these days myself, but now that the cloud is here, this type of tragedy can easily be avoided.
Where to go for some more serverless reading… 📚
Chris Feist wrote a post called Making serverless variables work for you to accompany his new
serverless-plugin-composed-vars plugin for the Serverless framework. I do this a bit differently, but this looks like a handy plugin.
Migrating a Serverless application backend to the Serverless Framework highlights Tai Nguyen Bui’s journey moving away from the console and into the world of serverless deployment automation.
Speaking of serverless journeys, How I Got Comfortable Building with Serverless highlights how Jun Fritz went from code bootcamp graduate, to Stackery employee, to confident serverless builder in just a few months. There is still much to learn, but it is fascinating how quickly people can get things up and running.
The state of serverless: 6 trends to watch highlights a fairly obvious (IMO) evolution of any new technology. However, I think that betting Knative will drive standardization is a bit off. We can argue about what serverless means all day long, but with CloudFlare workers moving compute to the edge, and AWS loading VMs closer to the metal with Firecracker, I personally see anything that adds more layers of abstraction to ephemeral functions being a step in the wrong direction. Maybe it’s just me.
In Dear Go — Thank You For Teaching Me PHP Was A Waste of My Time, Vern Keenan is pretty harsh about the future prospects of PHP. Not sure I agree with him on that, but he does make some good points about Go potentially becoming the dominant serverless runtime.
And finally, Zac Charles asks, What happens to running threads when a Lambda finishes executing? If you’re interested in the inner workings of Lambda functions and container reuse, give this short article a read.
When you’re curious what AWS has been working on… ☁️
There were a lot of serverless announcements and innovations at AWS over the last few months. If you’re having a hard time keeping up, take a look at Eric Johnson’s full recap: ICYMI: Serverless Q4 2018
The new AWS Backup lets you automate and centrally manage your backups across AWS services. Jerry Hargrove (aka @awsgeek) wasted no time putting together a cloud diagram for you. He’s also got a great one for the new Amazon DocumentDB service as well.
AWS also added S3 as a deployment action provider in CodePipeline. Check out this tutorial to learn how to Create a Pipeline That Uses Amazon S3 as a Deployment Provider. Plenty of cool use cases with this.
Two weeks ago AWS announced that AWS Step Functions would support resource tagging. Now they’re getting their very own Service Level Agreement with three 9s.
Serverless Star of the Week ⭐️
There is a very long list of people that 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 let me know.
This week’s star is Mikhail Shilkov (@MikhailShilkov). Mikhail is a Microsoft Azure MVP, a frequent conference speaker, and an advocate for all things serverless. His blog is loaded with insanely thorough articles about serverless (and functional programming) that are sure to help you level up your own skills. He mostly focuses on Microsoft, but has articles like this and this that can give you some much needed perspective in the overall serverless ecosystem. And today is his birthday, so Happy Birthday, Mikhail, and thanks for what you do! 🎂🎉🎈
Final Thoughts 🤔
Thank you for all the responses from last week. Everyone that sent me a message said they like the length and that they found it easy to skim and pick out the articles they were interested in. I’m glad you all like it. If you have any other thoughts, I’d be happy to hear them.
I hope you enjoyed this issue of Off-by-none. I love hearing your feedback and suggestions, it helps me make this newsletter better. Feel free to contact me via Twitter, LinkedIn, Facebook, or email and let me know your thoughts, criticisms, or how you’d like to contribute to Off-by-none.