Good Morning!

I’m at Google Cloud Next for at least part of this week; if you’re around and see me rushing around, please say hello.


I’m performing a live Q&A next month; submit your questions here!

From the Community

I miss fun things like this on Twitter post-API changes. Fortunately Mastodon steps up with AWS Status Poems.

Grafana has a post about scaling a memcached cluster to 50TB and from sheer budget alone I know they aren’t doing it on Elasticache.

Ooh, using the CDK to bridge GitHub Actions and Slack means I might do more with those two soon.

AWS legend James Hamilton reflects on Amazon Elastic Block Store at 15 Years. Holy… 100 trillion operations a day? 13 exabytes transferred daily? There’s scale and then there’s this.

El Reg asks whether we can trust ‘open source’ companies. I think the growing realization is that companies, when faced with hard times, will (however reluctantly) change the terms of the implicit bargain they make with their user base. Cory [sic] Doctrow calls this "enshittification" and I think the term fits perfectly.


Last Week In AWS: SageMaker Podcast HealthOmics

Last Week In AWS: Storing Logs You Never Read

Screaming in the Cloud: Cloud Compliance and the Ethics of AI with Levi McCormick

Screaming in the Cloud: The Value of Good Editing in Content Creation with Alysha Love

Choice Cuts

Amazon Aurora Global Database introduces Global Database Failover – Next new feature: "Automatic Global Database Failover." Next new feature after that: "Opt out of Automatic Global Database Failback" after someone learns just what "flapping" means as two out-of-sync systems vie desperately for control.

Amazon ElastiCache for Memcached simplifies creating new clusters in the AWS Management Console – This is great news for customers and an absolute disaster for any creators or educators who built "how to use ElastiCache" video content or included screenshots that now all need to be redone.

Improvements to multi-account management for Amazon GuardDuty – Credit where due: I like the idea of being able to do more and more from delegated administrator accounts; someday the only thing we’ll have left to do in the management account is fret about the bill.

AWS Certificate Manager introduces Enterprise Controls to help govern certificate issuance – I desperately want to know who took shadow IT to the next level and spun up an entire shadow corporate division. That’s my head-canon around why this change exists.

AWS Cost Explorer announces support for AWS Billing Conductor – It’s always good when different services within AWS Billing discover one another.

AWS Microservice Extractor now supports visualizing very large enterprise applications – How had I not heard of this before (because it’s a .NET thing, presumably)? The “wallet extractor” jokes just write themselves.

AWS re:Post launches an enhanced search experience – re:Post remains a wasteland where questions go to remain unanswered. Easy example of this: a little-known (to many; if you know this bully for you) bit of RDS trivia is that you can only stop an instance for 7 days; after that it returns to life like a zombie who whacked the snooze button too many times. Pretend you don’t know why this is, and see how carefully you have to craft a query in re:Post’s search to get this fairly common answer presented to you as the first search result. My best result so far has been "result number 3."

Announcing AWS ROSA console support for the ROSA with hosted control planes preview – This relocates the ROSA control plane from where it lives now (your AWS account / the Stone Age) into an AWS account controlled by Red Hat, which means that any issues are going to result in a three-way blamefest. Fire up your conference bridges / old timey party lines!

EC2 Hibernate now supports Amazon EC2 M7i and M7i-flex instances – Wait. Wait wait wait. Are you telling me that EC2 instances can become generally available without supporting the full EC2 feature-set? That is TERRIFYING. How is this even possible?

Manage Cost Allocation Tags with Last-Updated and Last-Used timestamps – Oh my stars thank you. "Do we or have we ever used this cost allocation tag" is a frequent question I encounter on client engagements.

Protecting an AWS Lambda function URL with Amazon CloudFront and Lambda@Edge – I prefer to do this via CloudFlare once you get the magic incantation correct, just because it’s so much nicer to work with.

Choose AWS Graviton and cloud storage for your Ethereum nodes infrastructure on AWS – Tainting the very solid Graviton and storage offerings that AWS has with the stink of blockchain is doing those services remarkably dirty.

How Amazon Finance Technologies built an event-driven and scalable remittance service using Amazon DynamoDB – I sent this article to my CFO so he could enjoy a particularly crappy episode of How It’s Made.

Upgrade from Amazon Aurora Serverless v1 to v2 with minimal downtime – Back when Aurora Serverless v2 came out, I reviewed it and was rather clear in my conclusion: these were basically two different services that behaved very differently. Has that clarified to a point where going from v1 to v2 is a clear upgrade?

Next Big Things for Retail – Generative AI leads the pack but isn’t alone – The next big thing for retail is likely the same that it’s been for a decade: Amazon is coming to kill your business. Now please give money to AWS.

Explain medical decisions in clinical settings using Amazon SageMaker Clarify – AI is generally unexplainable. The idea of using it for medical decisions and then asking it to explain itself is terrifying to me–but I’m not a doctor. I’m very, very curious to know what actual medical professionals think of this approach, because to me it’s dystopian in the extreme.

Build a serverless store finder site using Amazon Location Service – It sure would be nice if the default use case for so many services didn’t all revolve around "customers buying things." Yes, I get it, yay capitalism–but isn’t there a better narrative available to an enterprising cloud storyteller?

Configuring client IP address preservation with a Network Load Balancer in AWS Global Accelerator – This is a neat feature; historically you had to do all kinds of unpleasant hacks to achieve this sort of functionality.

How to use pulse-level control on OQC’s superconducting quantum computer – I’m not saying that this is a completely different field than the ones we normally work with in cloud computing, but I’ve read dissertations that were more approachable than this blog post.

AWS Digital Sovereignty Pledge: Announcing new dedicated infrastructure options – This is a subtle and remarkably genius play by AWS; it feels like it’s an end run around the growing realization by governments that they can shake down cloud providers for region-sized investments by passing data residency laws. Good for AWS; I have little patience for that sort of tomfoolery.


airflowctl helps you manage Apache Airflow projects, and this analysis gets into exactly how it does that.

… 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.