Welcome to issue number 76 of Last Week in AWS.
Happy Monday; I’m on a plane.
Are you in Birmingham, Alabama? Tomorrow (Tuesday) I’m giving a talk at the Birmingham AWS User Group called The Collected ‘Wisdom’ of Corey Quinn about modern AWS architectural best practices. You should come. I’m going to stop writing a newsletter intro and get that talk built instead.
The amount of data (and infrastructure) you have to work with is only going to keep increasing, and you need to both generate and synthesize an ever-increasing volume of logging data. This free eBook from Scalyr will help you do just that: look at what to log and how to log it with step-by-step guides for logging Java, C++, Python, C#, JS, Ruby, Go, Node.js, and whatever the hell Spring Boot is. Download now on the eBook format of your choice (Kindle, Google Play, iTunes, Nook, and Kobo). It’s impressively researched, and extremely timely. Thanks to Scalyr for their support of this ridiculous newsletter.
A video of my Observerless talk was posted from ServerlessConf.
An excellent read about the growing sentiment that multi-cloud is a trap.
The Containers on AWS site gives handy architecture patterns for ECS and Fargate.
If you’re hosting static sites, you probably want to look into automating it with CloudFormation, albeit begrudgingly.
While not strictly AWS focused, I wrote about the impact that Google EOLing Inbox has on perceptions of GCP.
CloudForecast talks about how they built their site with Serverless, and saved over a grand a month along the way.
I was interviewed on A Cloud Guru’s Serverless Superheroes feature– see my screed on why the cloud scales faster than your budget.
A dive through hosting Python web applications with AWS Lambda and Flask.
If you’re using CloudFlare as a part of your environment in an acknowledgement that CloudFront is… not great, you may want to consider automating DNS updates via Lambda.
It’s a sad fact of life that Forbes has done such a bang-up job of diluting its own brand with complete crap by “Contributors” that I really had to dig into it to figure out if their “the Cloud 100” was meaningful, or a waste of bandwidth.
GorillaStack has an interesting take on splitting your serverless framework API with AWS.
The site feels a bit salesy, but these Redshift performance tuning techniques are solid. Articles like this always leave me feeling conflicted…
This is a bit meta– a post that distills down a talk on the topic of Apache Kafka with data lakes.
The joy of microservices means that things break and you’re left solving a murder mystery. GoCD has a fascinating post on how to remediate failures with microservicesthat everyone assures you couldn’t possibly happen when they’re cramming microservices based architectures down your throat. Check it out– let me know what you think. Thanks to ThoughtWorks for their ongoing support of this newsletter.
Choice Cuts From the AWS Blog
Amazon Neptune Achieves HIPAA Eligibility – The Giraffe Database can now be used for vetinary records.
Introducing the AWS Cloud Management Tools Competency – There are some familiar names here. Behold the list of companies that Amazon will surely do its level best to crush by 2020…
AWS Lambda Supports PowerShell Core – Not Ruby. PowerShell. Not Rust. PowerShell. Not PHP. PowerShell. Not bash. PowerShell.
AWS Serverless Navigate Program – This wonderful program explains what the heck Serverless is to Amazon partners. “Here, do business with a partner we had to explain Serverless to” is going to be an interesting sales pitch for Amazon reps…
Introducing AWS Systems Manager Session Manager – I’m going to write a system to manage it, and then give a talk about it at re:Invent. Therefore, if you have trouble finding a seat, you can talk to the AWS Systems Manager Session Manager Management System session manager. Honestly, is the person who named this service okay, medically speaking? We can start a crowdfunding campaign if you need one.
Introducing private registry authentication support for AWS Fargate – Instead of being limited solely to ECR, you can now additionally use other private registries or repositories that actually work.
Classifying high-resolution chest x-ray medical images with Amazon SageMaker | AWS Machine Learning Blog – To turn off the snark for a minute, when I was a kid I had some rather serious health issues that are thankfully decades behind me. In my early teens I had more chest x-rays than I can count. Interpreting a thoracic x-ray is incredibly complex, and requires non-trivial training, even for very specific diagnostic work. If work like this reduces the human diagnostic burden, I’m all for it. Be forewarned– this article reads like two PhDs wrote it– which they did. “Approachable” it is not.
The biggest pain for Kubernetes is that it magically tends to assume that your applications already support it. That’s obnoxious– most of them don’t. To that end, DigitalOcean has a fantastic guide on modernizing applications for Kubernetes that just came out. I’m a fan! Thanks to them, both for their support of this newsletter as well as letting me pet one of their dogs last week.
This handy script uploads EC2 logs to S3 on shutdown, which is helpful when you’d like to know why the ASG is whacking your nodes.
This is is such a good idea I’m surprised it hasn’t been done before. awslog shows the changes to AWS resources over time.
There are a lot of Slack bots that estimate your AWS charges, but AWSEstimatedCharges2Slack is open source and doesn’t require you to trust a third party. Y’know… other than Slack.
…and that’s what happened Last Week in AWS.