Good morning!

Welcome to issue number 92 of Last Week in AWS.

Last week this newsletter broke the 10,000 subscriber mark for the first time. I’m stunned that this many of you are reading; at times it doesn’t seem quite real. For example, in a typical week I’ll get 4-6 responses (not including auto-responders; good lord) from the list, and another couple of pings on Twitter. It’s surreal to think of how many people that is. That’s “needs a stadium to hold them all.” That’s “almost a quarter of re:Invent 2018 attendees.” That’s “over twice the number of AWS services with ridiculous names.”

So thank you all for reading. I appreciate your support, your sharing this with your networks, your willingness to respond to the nonsense I throw out every Monday. Here’s to the next 10,000.

This week’s issue is sponsored by Scalyr.

Christian Meléndez (Cloud Architect at Equinix) and Dave McAllister (Scalyr Community Guy) are hosting an online workshop (Tue 1/29 at 9am PT) on developing and monitoring Kubernetes. They’ll be showing live code and command examples to accompany the concepts we’re explaining, and share some of the common challenges of monitoring activity on Kubernetes. RSVP here.

Community Contributions

A discussion of why Serverless is taking over the world. There’s a lot that can be said about Serverless, and will be said in the coming months. I have a few blog posts that are being sent for editing, but this is a business revolution more-so than a technical one.

Of all the ways to save money when running Kubernetes on AWS, it’s hard to imagine any of them irking AWS more than Classic ELBs.

The idea of using Lambda layers to inject chaos is a fascinating one. Expect more development around ideas like this one over the coming months.

A pair of posts this week about Lambda. The first discusses what happens to running threads when a Lambda finishes executing. The second discusses the fact that files survive Lambda timeouts, though your handler doesn’t.

I appeared on Software Engineering Daily to discuss AWS analysis. I think it came out rather well! It’s always a treat and an honor to be invited onto podcasts, conference stages, and other platforms to share my thoughts.

Kevin Kuchta made a comic for me. Be forewarned, there’s a bit of salty language.

If you’ve applied for a job with First National Real Estate in Australia, there’s a great chance that you’ve just been the victim of this week’s S3 Bucket Negligence Award. Might I suggest applying to work somewhere more competent?

I was impressed enough by Console Recorder a few weeks back that I cornered its creator Ian McKay on a podcast; Screaming in the Cloud Episode 44: Disagree in Commits.


It’s time to bring back the Jobs section. Let’s try something new. Job posts here are free–unless you’d like me to use your copy, in which case I will charge you. Consider it an experiment with an extortion-based business model!

I also have someone interesting to tell you about. He’s passionate about cloud engineering, and has a background in writing Kotlin and React. Picture him as someone who can help write effective tooling for your cloud efforts. He’s relatively new to the industry, but you wouldn’t know it to talk with him. He asks the right questions–and he has my endorsement. If someone like this fits into your business, hit reply; I’d be pleased to introduce you to him. Note that he’s currently looking to remain in Las Vegas, so if you’re not there, remote is going to have to be something your company is comfortable with.

Choice Cuts From the AWS Blog

Announcing Windows Server 2019 AMIs for Amazon EC2 – Two weeks in to 2019 and we’re already seeing it appear in product names. This also serves as your reminder that “AMI” has three syllables.

Introducing Amazon DocumentDB (with MongoDB compatibility) – Generally available – As soon as this service was announced my mentions lit up with people asking me to weigh in on its name. Sorry to disappoint; when your biggest competitor is named “Mongo,” you pretty much get a free pass to call the service whatever the hell you want. Cue a thousand takes about how MongoDB is doomed–somehow I don’t see it happening.

Amazon EC2 Spot now Provides Results for the “Describe Spot Instance Request” in Multiple Pages – If you wrote a script to manage Spot instances that broke over the past week, here’s why. Enjoy the pagination dance!

Amazon EMR Announces 99.9% Service Level Agreement – Someone at AWS apparently woke up and realized that large enterprises have “SLA” as a checkbox item for service uptake. That was a legendary nap that spanned at least a decade…

Amazon QuickSight Launches Pivot Table Enhancements, Cross-Schema Joins and More – AWS Employee: “Wow, QuickSight is terrible. I pity whichever company built it.” 
Other, vastly more senior AWS Employee: “You know that that’s our product, right?” 
Suddenly QuickSight is seeing a bunch of improvements for no discernable reason…

Announcing AWS Fargate Price Reduction by Up To 50% – It just got a lot cheaper to run containerized workloads on Fargate. There are still barriers to adoption for a lot of workload, but “it costs Splunk money” is no longer first and foremost.

AWS Single Sign-On Now Enables You to Direct Users to a Specific AWS Management Console Page – This is a welcome change. I can’t wait for aws-vault to support it.

How to centralize and automate IAM policy creation in sandbox, development, and test environments | AWS Security Blog – A lot of these solutions bolt together twenty different services, but this one’s compact enough that it seems appropriate to mention. Let me know if you’d like me to include more or less content like this one; I’m here for you folks!

New AWS services launch with HIPAA, PCI, ISO, and SOC – a company first | AWS Security Blog – I’m a fan of the end result, but my brain boggles at the sheer scale of the amount of work it takes to launch a service with just one compliance certification–let alone four of them. I can’t even snark about this. It’s just an incredible feat of execution…


Want to pick an EC2 instance, but are lost within AWS’s constant assault on the alphabet when it comes to instance types? is here to solve that one for you.

A set of tooling for repeatable Lambda benchmarking. This is handy for figuring out various failure modes and degradation patterns for your serverless applications.

…and that’s what happened Last Week in AWS.

Newsletter Footer

Sign up for Last Week in AWS

Stay up to date on the latest AWS news, opinions, and tools, all lovingly sprinkled with a bit of snark.

"*" indicates required fields

This field is for validation purposes and should be left unchanged.
Sponsor Icon Footer

Sponsor a Newsletter Issue

Reach over 30,000 discerning engineers, managers, and enthusiasts who actually care about the state of Amazon’s cloud ecosystems.