Good morning!

Welcome to issue number 58 of Last Week in AWS.

Last week I got to visit Seattle for Microsoft Build. Azure is… well, it’s no AWS, but that’s not really its fault. I did get to speak with the excellently named Corey Sanders, Corporate VP of Azure Compute while I was there. I shoved a microphone in his face; check out Cloud Coreyography.This week’s issue is sponsored by DigitalOcean:

Digital Ocean Logo

Looking for a high-performance cloud platform with no billing surprises? Always know what you’ll pay each month on DigitalOcean with monthly caps and flat pricing across regions. Get started with a $100 credit.

Community Contributions

As AWS readies the AWS/400 it must surely be constructing, this article helps put us into the mindset of the mainframe managers who are just now approaching cloud– Demystifying Cloud for Mainframe Managers.

I missed this when it came out a few months ago. There were some changes a year or three back to make this globally consistent, so it’s worth reviewing the S3 Consistency Model if you have a minute.

Honeycomb has a public postmortem about their experience with RDS pooping itself. They of course took a slightly more professional approach to it than I did.

Continuous delivery of a static website is now more feasible due to GitHub webhook support.

This well written reddit post goes into depth on where to get started if you’ve taken a blow to the head and thus want to be a cloud engineer.

And now, the first of many re:Invent guides for 2018. I’ll see some of you there.

Xero goes into depth about the impact that their massive AWS migration had upon their business.

LucidChart, a great program that I recommend, discusses their database migration to Aurora. I’d love to chat with the folks there about a few things the article doesn’t get into depth with, if any of you know someone I can talk to…

“What do you mean we pay per unattached elastic IP address?!”


The bill fights between AWS and RyanAir must be the stuff of freaking LEGEND now that the latter has selected the former as its cloud provider of choice.

Choice Cuts From the AWS Blog

AWS SAM CLI Simplifies Building Serverless Apps with the SAM init Command – The opinionated view of the Serverless world now has a whole new command line syntax. I’m sure this won’t confuse anybody.

Application Load Balancer and Network Load Balancer now Support Resource- and Tag-based Permissions – Rejoice, for you can now explicitly route traffic to instances tagged as “not crappy.”

Amazon Elasticsearch Service Offers Additional Cost Savings with Reserved Instances – The search service that doesn’t work very well now gets reserved instance treatment, adding the complexity that nobody outside of finance was asking for.

Amazon GuardDuty Adds Capability to Automatically Archive Findings – Good news– you can now pay AWS to automatically ignore security reports instead of doing it yourself manually the way you have been to date.

AWS Linked Account Bills now Reflect an Account’s Unblended Costs – If anyone would like to read this announcement and explain it to me, I’d appreciate it. More seriously, this is what I do for a living, and it still took me three reads to make heads or tails out of what this change does.

Get Latest Console Output on EC2 Instances – It doesn’t stand up to Azure’s “virtual serial console” feature for one. For another… this is designed to be used outside of state transition events such as a start or stop. …I’m sorry, I have to ask: what the heck terrible workflow is this empowering?! “If there are more than forty characters in the last line of console output” is a terrible conditional.

Introducing Optimize CPUs for Amazon EC2 Instances – “You can now disable CPUs at instance launch. We’re not going to charge you a penny less for doing that, but Oracle licensing is done on a per-core basis” says Amazon, firing a warning shot directly into Larry Ellison’s front door.

Introducing the IoT Device Simulator – I can’t want to simulate IoT Devices, but first I need to wait for AWS to release a Business Model Simulator so I can come up with a viable use case for it.

Monitor your Reserved Instance coverage by receiving alerts via AWS Budgets – “It’s 3AM. Do you know what your instance coverage is?” Last week’s news story about the woman who sent 65,000 text messages to someone she’d been on a single date with now comes to AWS.

Stream Real-Time Data in Apache Parquet or ORC Format Using Amazon Kinesis Data Firehose – This is fascinating; Kinesis Data Firehose can transform data formats on the fly. I believe this is the first time we’ve seen this behavior outside of DMS and Glue…

Amazon Aurora Backtrack – Turn Back Time | AWS News Blog – ♪ ♫ ♬ If I could turn back time

If I weren’t a crap DBA

I’d take back those ALTER TABLE deploys

And not trash my SLA

I don’t know why I pushed the change I did

I don’t know why it made sense in my head

DROP’s like a knife they can cut deep inside

Reindexes are weapons, they can trash several 9’s

I didn’t really mean do that

I didn’t want to get let go

I know I got you paged, but hey boss…

♪ ♫ ♬

If I could turn back time

If I were a DBA

I’d roll back those changes that broke things

In under a day

If I could ditch my VARs

I’d save a hundred grand or two

Then you’d not sack me, sack me, like you’re about to

If I could turn back time…

♪ ♫ ♬

My cloud doesn’t love me, and Container CI/CD | Dear DevOps Abby – The first Ask AWS Abby kicks off with a pair of excellent questions. “What, Corey isn’t going to snark about the questions?” I most certainly am not– there’s no such thing as a dumb question, and anyone who tells you otherwise is garbage.

New Low-Cost Option to Access AWS Direct Connect through Internet2 | AWS Government, Education, & Nonprofits Blog – For those institutions who are part of internet2, this is worth paying attention to. I’m personally holding out for internet5; it’s likely to have fixed a lot of the problems by then.


A Lambda function that archives other Lambda functions is boring, but I love the name of SmushIt.

I’m surprised this isn’t native; Surge does multipart uploads and downloads to Glacier.

A bit dated, but cloudfrunt helps you identify misconfigured CloudFront domains.

I’m somewhat skeptical that aws_public_ips can fetch all public IP addresses tied to your account, but I’d give it a shot if I needed such a thing.

If you’re working with Terraform and want to do some automated testing, Terratest is a go library that may help out.

Sked is a Python utility to make working with scheduled instances a lot easier.

Terrible Tip of the Week

…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.