Welcome to issue number 77 of Last Week in AWS.
Some noise was made this week by a Citi analyst about the idea of splitting Amazon into two companies. A writer at the Motley Fool is convinced that this absolutely won’t happen, but I’m generally hesitant to categorically state that Amazon will or won’t do something. I do know it’d be awfully nice on my end to be able to focus on AWS changes without having to worry about privacy issues from their talking toilet cameras or whatever it is that Amazon released last week at their Alexa event…
If you’re a GoCD user, you might be spending time working around a lack of visibility into your pipeline. If so, you’re likely not aware of their analytics plugin. You can’t improve what you don’t measure; give it a spin. Thanks to ThoughtWorks for their support of this newsletter.
A recurring legend that’s almost impossible to kill keeps coming up in conversation– the idea that AWS came from Amazon’s spare capacity. One of the builders of EC2 sets the record straight.
High Scalability has an interesting point to make– for the past decade AWS’s security model has been unreasonably effective. This delves deeply into why that is.
I love a good metaphor– this one’s equating medieval cities to cloud security.
Another article on boosting the delivery of static websites. I’ll be tweaking my own workflow shortly.
Cloudonaut returns with a discovery-dive into Cloud Directory, a neglected serverless data store.
A discussion into the realities today of Serverless multi-cloud.
A discussion of the mindset change that migrating to the cloud requires for compliance and audit folks.
A guide into using custom build environments to speed up AWS CodeBuild. I will say, when properly configured it’s not a slow service at all.
Ben Kehoe (iRobot’s Cloud Robotic Research Scientist) appears on a podcast to discuss their use of serverless.
I caught up with Rowan Udell of Versent to talk about serverless patterns in Screaming in the Cloud Episode 28: Serverless as a Consulting Cash Register (now accepting Bitcoin!).
While not quite an S3 Bucket Negligence Award, Veeam left 445 million customer records exposed on AWS via what looks to be a misconfigured MongoDB server, and probably triggering a GDPR investigation. Smooth.
The amount of data (and infrastructure) you have to work with only keeps increasing – and you have to generate logs from all of it and synthesize them as much as possible. This free eBook from Scalyr will help you do just that: figure out what to log and how to log it. You’ll get step-by-step guides for logging in Java, C++, Python, C#, JS, Ruby, Go, Node.js, and whatever the hell Spring Boot is. Download 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.
Choice Cuts From the AWS Blog
Amazon Aurora Parallel Query is Generally Available – While interesting, I’d be reluctant to architect an application around a capability that only one large vendor is able to provide,
Amazon EKS Simplifies Cluster Setup with update-kubeconfig CLI Command – Y’know what’s even simpler? Avoiding the tire fire that is current-generation EKS altogether. There’s something golden buried underneath what it currently is– but it’s not here yet.
Amazon ElastiCache for Redis adds read replica scaling for Redis Cluster – As the war between Redis and AWS escalates, AWS adds capabilities that you’ve had by running your own Redis clusters for ages now…
Amazon S3 Announces Selective Cross-Region Replication Based on Object Tags – If your S3 bucket organization is a complete mess, but by some miracle your tags aren’t, you can now replicate objects on that basis alone.
Amazon Elasticsearch Service now supports encrypted communication between Elasticsearch nodes – “Wait, you mean it didn’t before?” This is one of those things people always sort of assumed was being done already.
AWS Database Migration Service Makes It Easier to Migrate Cassandra Databases to Amazon DynamoDB – You can now swap one form of noSQL misery for another.
AWS Storage Gateway Is Now Available as a Hardware Appliance – If you want to order yourself an appliance from Amazon that they’ll run inside of your datacenter for you, you can now buy one. This is nifty, albeit expensive for my purposes.
Save time and money by filtering faces during indexing with Amazon Rekognition – Now Rekognition can determine whether a face is a hot dog before it starts identifying specifics.
AWS CLI v2 development | AWS Developer Blog – The next generation of the AWS CLI is in development, and they apparently want feedback through a formal process that isn’t “accosting employees on Twitter.”
AWS Organizations now requires email address verification in order to invite accounts to an organization | AWS Security Blog – I’d love to know the backstory behind this one. What did someone do that required this feature being rolled out?
マンガで学ぶ AWS 『なな転び八起の AWS 開発日記』| AWS – AWS Japan has apparently authored a manga comic about their offerings. Why can’t AWS America have this sense of whimsy to its documentation?
This week the head of UX at DigitalOcean’s say down with the folks at Invision’s Inside Design blog to discuss five best practices when you’re designing for developers. As someone who develops mostly on the command line, I found this rather helpful– and it shows in how DigitalOcean tends to think about user experience. Thanks to DigitalOcean for their continuing support.
If you need a library to tell you what cloud provider code is running on, NOAA is a good start.
Have a Lambda function that runs a daily cost report and spits it out to Slack on a daily basis.
A great list of health check examples can be found in the AWS Health Tools repository. This is great for monitoring system integration.
…and that’s what happened Last Week in AWS.