Welcome to issue number 50 of Last Week in AWS.
Last weekend at SCaLE I gave a 5 minute Ignite talk (20 slides that auto-advance every 15 seconds) called “Death of an Enterprise Software Salesman.” I used these five minutes to sell the audience something I can neither describe nor define. Please feel free to watch it, and fall under my salesy spell.
The BBC reports that a variety of firms / do-gooders are lashing out at insecure S3 buckets they discover– via placing text files into them warning about the risk. I can confirm this myself based upon a few insecure buckets I’ve tripped over in recent weeks that have already been tagged with helpful files explaining the issue.
While not strictly AWS oriented or bitingly sarcastic, if you’re not subscribed to the weekly DevOps’ish newsletter you’re missing out on some great content– such as opinionated thinkpieces declaring that Docker, Inc is Dead, a DevOps reading list, and a well researched on-ramp for Kubernetes.
Cloudonaut returns with an exploration of why “observerless” approaches to t2 instances is a bad idea; the burst credits need to be monitored. If your non-t2 instances slam to a halt for no reason, it’s probably some other reason.
A great CTF retro explaining a number of AWS and Git concepts. Well worth the read; mind the lessons embedded within.
Pour one out for Evident.io; Palo Alto Networks has entered into an agreement to acquire them. This could end well, but likely won’t.
A step-by-step tutorial to building a personal journal web app is interesting. That said, I prefer to keep my thoughts in an email, which I then send to all of you every week for safekeeping.
A thoughtful piece on whether or not Kubernetes should be scared of AWS. I don’t know as an open source project should be “scared,” exactly…
Seven quick tips about AWS Lambda. A couple of these were new to me.
A fun exploration of building a serverless CMS. There are a lot of static site generators the author could have picked, but take a look at the victor: Hugo.
In my days as a hiring manager for SRE teams, I used to ask a technical interview question that distilled down to “tell me how you’d build a URL shortener service.” The problem was then complicated with further requirements (make it scale, build a HA story, survive the loss of a datacenter without going split-brain) added in. This URL shortener built with Go and Lambda demo solves for all of them, very neatly.
This week’s S3 Bucket Negligence Award is brought to you by Walmart— or more accurately, one of their jewelry partners. For a company that made a lot of noise last year by asking that all of their vendors move off of AWS entirely, this is something of a black eye.
AWSgeek Jerry Hargrove has two posts this week; one exploring SageMaker, and the other exploring GuardDuty. As always, his diagrams are informative, beautiful, and have just enough whimsy to make me smile without losing the thread of what he’s conveying.
VMware is planning to let customers vMotion instances between regions. This solves a few architectural hurdles neatly, at the cost of having VMware still playing a pivotal role within your infrastructure in 2018.
Choice Cuts From the AWS Blog
Amazon Chime Introduces Usage-Based Frustration-Free Pricing – Chime (“like Slack, only worse and almost nobody uses it”) now gets an additional “flexible” billing model that’s much more similar to Slack’s (“only more expensive, and almost nobody will use it”). To be fair, it’s still better than IRC or email…
Join the AWS Quest – Help me to Rebuild Ozz! – Jeff Barr’s robot has exploded, like a monolithic application running in a single AZ within us-east-1. Help him find the pieces to put it back together again.
Amazon Elastic File System (Amazon EFS) Available in US West (Northern California) Region – The more-expensive, reportedly power-constrained Northern California region now has NFS-as-a-Service, enabling yet another layer of poor decisions in some shops.
Announcing Gluon Support for AWS DeepLens – This may be the only time in history where the phrases “glue-on” and “expensive camera” led to a positive outcome.
AWS CloudTrail Log Search Using Amazon Athena – “What if you could search your logs at rest in S3” is much better spin than “we admit the current state of log analysis with CloudTrail is a confusing disaster, and we’re working diligently to fix it.” Snark aside, this is a welcome change.
AWS Serverless Application Model (AWS SAM) Supports Additional Amazon API Gateway Features – This is a great list of features that SAM supports– yet I’m vaguely horrified that it didn’t support them before.
CloudFront now Supports ECDSA Certificates for HTTPS Connections to Origins – “Hey Corey, this is big news; how would you explain it to a relatively non-technical user?” “I’m not sure– can you explain it to me first?”
New AMIs Available for Testing Longer Format Image IDs – I’m mildly irked by how Amazon can invest the time and resources into building entire AMIs to support longer format image IDs, but not the extra time into pronouncing “AMI” properly with all three syllables.
Manage your Amazon RDS, Redshift, and ElastiCache Reservations via the Reserved Instance (RI) Coverage Report in AWS Cost Explorer – The RI coverage reports get a lot better coverage, as you get to begin the fun process of analyzing exactly what you predict your three year growth of Redis is likely to be.
Our Newest AWS Community Heroes (Spring 2018 Edition) | AWS News Blog – AWS announces “five new community heroes” at time of publication, but only gives biographies for four of them. I blame fencepost errors. I’m glad I was able to sandwich this update in.
Predict March Madness using Amazon Sagemaker | AWS Machine Learning Blog – While fascinating in theory, if this worked reliably in practice whoever discovered it would immediately stop talking, quit their job, and shortly retire to their own private island.
An EBS autotagger that’s packaged in Terraform and ready to go. I’d love to see this expanded beyond just autoscaling groups; there’s a very real need here.
A good “learn by example” application/tutorial that gets into running Go applications in Lambda. I’m a sucker for this kind of thing.
…and that’s what happened Last Week in AWS.