Last Week in AWS Logo

Good morning!

Welcome to issue number 81 of Last Week in AWS.

This year's charity t-shirt is now available for sale--this week only! All proceeds benefit St. Jude Children's Research Hospital.

If this newsletter has been at all helpful, entertaining, or you're against childhood cancer, please consider buying one.

Should you not like the shirt, you can also make a donation through the link above.

With over 150 instance types available in us-east-1 alone (not to mention tenancy, standard vs convertible, zone specific or not, no/partial/full upfront variables), the process of buying Reserved Instances is the clearest sign yet that AWS hates you. Here to bring sanity to the chaos is Cloudability with a guide to reserved instances that helps you get it right the first time. Measure twice, buy once. Thanks to Cloudability for their support of this newsletter.

Community Contributions

Perhaps you didn't read the introduction to this issue. Perhaps you're an AWS employee who worries that you might suffer unpleasant consequences for buying a snarky shirt. You should buy one; the Amazon Leadership Principles command it!

Adroll talks about how they used EC2 Spot Fleet to save a pile of money on their AWS bill.

A great set of benchmarks on comparative performance of gp2 volumes vs NVME. I'd been wondering about this myself...

Cloudonaut returns, with a post on customizing rate limiting for API Gateway.

I'm not at all suggesting that CloudFormation needs some tooling to make it usable, but here's cloudformation-plus. It makes CloudFormation a lot more usable, and it's made by HPE.

This is fascinating; AWS has dropped pre-requisites for their certification exams. What this means is that there's nothing stopping me from walking in and failing the Networking Speciality AWS exam, despite the fact that I only have the Cloud Practitioner certification. Historically I'd have had to get at least two more certs to unlock the chance to embarrass myself on that test.

AWS Advent is looking for contributors. The world's heard enough from me; someone else should write for them this year!

A handy list of 10 AppSync features you didn't know about; "what even is it" didn't make the list.

A beginner's guide to working with S3 via Boto3 and Python.

Colm MacCárthaigh, who has forgotten more about networking, security, and cryptography than the rest of us will ever know, had a brilliant Twitter thread (thread reader link) that explained how Amazon avoids last week's libssh problem.

Jerry Hargrove (AWSgeek) has outdone himself this time-- his visual service summary of the enormity of S3 is a masterwork, and (as you can probably guess) enormous. I want this as a poster...

Not quite an S3 Bucket Negligence Award, an abandoned application had its S3 bucket hijacked and used to serve malware. Amazing; defunct S3 buckets can be re-registered. I don't know as I've ever seen this before.

Put on your tin-foil hats; the Pentagon and a Tea Party PAC are collaborating. Not on anything nefarious or subversive, but rather in sharing this week's S3 Bucket Negligence Award. Bravo.

Jeremy Daly weighs in on what 15 minute Lambda functions tell us about the future of Serverless.

I got to speak with Brian Nelson, the founder of the Lambda School. Check out Episode 32 of Screaming in the Cloud: A New Approach to "Hire Ed.".

GoCD announces a raft of new features. Most relevant to readers of this newsletter are pipelines as code, vastly improved container-based workflows, AWS specific integrations, and more. Thanks again to ThoughtWorks (they make GoCD!) for their continuing support of this newsletter.


Friend of the newsletter Adroll is hiring for a senior infrastructure engineer. They're good folks who leverage a lot of what AWS has to offer, come up with innovative solutions and share them, and provide a retirement job for AWS Community Hero Valentino Volonghi, who as of this writing has worked there long enough for his tenure to get a driver's license.

Are you able to eat two entire pizzas by yourself? Are you capable of building amazing things and giving them terrible names when you show them to the world? Are you willing to mispronounce "AMI" with only two syllables? Consider working at AWS. If you already work at AWS, consider sponsoring this newsletter and replacing my copy with your own!

Are you annoyed with the current state of AWS providers in Terraform? Well stop complaining, go work at Hashicorp, and get paid to fix it yourself!

Choice Cuts From the AWS Blog

Amazon WorkSpaces introduces a new PowerPro bundle for resource intensive workloads - This was announced a day earlier in the official AWS support forums. For that day I was terrified that I was going to have to monitor still more release channels. It wasn't a good feeling! Fortunately, a new powerful WorkSpace is indeed a good feeling.

Announcing Managed Databases for Amazon Lightsail - It feels like they're really going back to first principles here and re-imagining EC2. I'm going to say it: I think this is a great thing. Expanding Lightsail's capability means that AWS becomes accessible to customers who until now haven't been willing to embrace the required complexity. Some of you may disagree with me on this one, and that's fine; given a choice between "simplicity of messaging" and "meeting customers where they are," Amazon biases for the latter every time, and there's validity to that decision.

AWS Lambda announces service level agreement - This is huge--not because anyone's going to ever get a large amount of money back from AWS due to an outage, but because it says "we're serious about this platform" in terms that large enterprise buyers understand. I've been asking for this for a while, and I'm thrilled that it's here.

Encrypt your previously unencrypted Amazon Redshift cluster with 1-click - There's someone reading this who just finished rebuilding their entire Redshift data warehouse last week who's about a second away from slamming their computer onto the ground and rage-quitting their job.

Performance Insights is Generally Available on Amazon RDS for MySQL 5.6 - Performance Insights has left Developer Preview, entered General Availability, broken down your office door, stormed into the DBA's cubicle, and rubbed their nose in their sins.

AWS Public Datasets Now Available from the German Meteorological Office, Broad Institute, Chan Zuckerberg Biohub,, and Others - I love the public dataset program, and I'm glad to see it expanding. While you're collecting enormous datasets for public analysis, can you maybe throw up a large sample AWS bill for me? It'd make some of my projects simplier if they weren't all NDA encumbered...

Specify Parameter Groups when Restoring Amazon RDS Backups - Your database restores just got a lot less painful.


You can manage S3 objects via CloudFormation with this template.

If you want to upload files to multiple clouds in parallel, consider cloudupload.

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

I’m Corey Quinn. I help people significantly reduce and understand their AWS bills and speak broadly on the conference circuit. In addition to this newsletter, I host the Screaming in the Cloud podcast about the business of cloud computing, featuring me talking to folks who are good at things; it's a nice contrast.

If you’ve enjoyed reading this, tell your friends to sign up at (or post a link in your company Slack team!) about it. As always, if you’ve seen a blog post, a tool, or anything else AWS related that you think the rest of the community should hear about, send them my way. You can either hit reply– or join the #lastweekinaws channel on the og-aws Slack team.

List archives are always available at