Welcome to issue 167 of Last Week in AWS.
Today is Apple Keynote day, so I look forward to seeing who drops bad news in a few hours in the hopes that it gets buried.
We also discovered a new AWS metric that we expect to start being billed upon shortly; more details below (and in the subject line, too).
From the Community
About 80% of software teams that do not yet practice observability plan to reach an advanced level within the next 2 years. We asked your peers to tell us how they’re doing.
Advanced observability teams are 3X more likely to work in an organization that understands the breadth and impact of their tech debt. A whopping 92% are confident they proactively notice and catch bugs after code is deployed to production.
If you don’t have MFA turned on for at least your admin level AWS accounts, stop reading and go fix that immediately.
A rundown on the new HTTP APIs vs. the old style API Gateways.
It’s always worth revisiting this series of python lambda decorators. It makes some common tasks way easier.
Ah, a use case emerges for Lambda with EFS that isn’t “shill for AWS’s machine learning nonsense.”
I love the idea of a Denial of Wallet attach, and AWS is the clear leader in these.
NextRoll / AdRoll / a roll of some kind talks about how they save money with Spot via some custom tooling.
Jerry has failed to capture the most important metrics in his visual summary of the AWS Snow Family. Let me fix it: the Snowmobile clocks in at 122 GB per ounce, the Snowball Edge (Storage Optimized) comes in at 100 GB per ounce, and the newly launched SnowCone hits 111 GB per ounce. I can’t wait to get billed on this exciting new dimension…
I didn’t realize AWS had seen a 2.3 Tbps DDoS attack that lasted days back in February.
Former AWS VP / distinguished engineer Tim Bray analyzes AWS’s Share of Amazon’s Profit. He gets some things wrong with regard to the financial analysis, but the underlying conclusion is that AWS’s profitability helps finance the retail arm to the point where “the cloud computing stuff is helping crush small businesses” becomes a credible narrative.
Another S3 Bucket Negligence Award, this time impacting over 30,000 sales agents in Italy.
TechRadar apparently believes that AWS believes it’s time to go all-in on cloud is newsworthy. Well… yes. The cloud provider would indeed believe that, in much the same way that I believe that now is the time to hurl your marketing budgets into this newsletter. Neither AWS nor I are particularly objective on our respective points.
Now that the goverment of the Philippines has mandated a cloud-first approach, a Manila AWS region seems like only a matter of time.
An S3 Bucket Negligence award has now exposed the intimate photos, chats, and more for customers of dating sites 3somes, Cougary, Gay Daddy Bear, Xpal, BBW Dating, Casualx, SugarD, Herpes Dating, and GHunt.
If you’ve got an interesting job for this newsletter’s eminently employable subscribers, get in touch!
If you’re looking for a senior management role, consider leading the Well Architected Tech Leads team at AWS. The Senior SA Manager, Well-Architected Tech Leads Leader will drive and improve best practices across a global team, helping customers use AWS better. (Let’s not kid ourselves; some of them are closer to the ideal cloud usage pattern than others, which is why Well Architected exists in the first place…) With roles in several states including California, this is a job of interest to some of you; check it out.
If you’re a Solutions Architect on the Well Architected team, you could slur your words slightly and be a Swell Architect. If that’s not enough to inspire you to greatness, consider the joy in helping customers and partners design better ways of working with the cloud, but not having to stick around for their terrible interpretation of what implementing that architecture looks like. If this sounds at all appealing, consider becoming a AWS SWell-Architected Solutions Architect. Several positions in several states are available; check them out.
Right now, the Well Architected Tool is pretty much a sad checklist. Amazon is looking for a systems application engineer to turn this into something great, since it turns out that after fifteen years of running public-facing web services, one key lesson is that computers are better at rote repetition than people are. It’s worth highlighting that this role asks for 0-3 years as a developer, so if you’re looking for a career change or breakthrough role, this might me of interest to you.
Sick of having to manage different CI servers and tools? Of course you are, even if you can’t admit it in public. Let’s talk about Buildkite; a unifying voice in this sea of madness. It has an easy-to-use web UI, extensive docs, and a portable agent that runs on any hardware or container runtime. You want to talk scale? Shopify has happily used Buildkite to grow from 300 to 3000 engineers–while keeping builds under 5 minutes. Check it out at buildkite.com. They’ve even got a CloudFormation stack if you want to cosplay as a responsible engineer!
Amazon Aurora Global Database supports read replica write forwarding – There are three data consistency models here that you can choose between depending upon your needs: session, global, or eventual. If you couldn’t possibly care less about data consistency, then the correct decision is of course to use MongoDB instead.
Amazon Corretto for Alpine Linux now in preview – Alpine is a Linux distro designed for security, simplicity, and resource efficiency. In the exact opposite direction from those design goals, Amazon now lets you throw a whole JVM into it.
Amazon EC2 Auto Scaling now supports Instance Refresh within Auto Scaling Groups – “Hey, replace all of the instances in my auto scaling group” is a neat feature, first proposed by the very first auto scaling group customer probably three minutes after launching their first auto scaling group in May of 2009.
Amazon ECS Capacity Providers Now Support Delete Functionality – “Uh, folks? The list of Capacity Providers in the console is now over eighty pages long and growing rapidly; can we please get around to finishing the API? Please?”
Amazon Pinpoint now supports International Long Distance Operator (ILDO) routes to send SMS messages to India – Now there’s a web server running on my phone and my cloud provider thinks it’s a telephone. Ma Bell, if only you’d lived to see this.
Amazon Polly launches a child US English NTTS Voice – If you think you want children, just wait until you hear one read your AWS bill to you.
Amazon RDS on VMware Adds Support for Read Replica – Amazon services on VMware are exciting, because they provide a sneak peak into what we’d all be dealing with if AWS had taken an alternate path and let us run their software in our own horrible datacenters.
Amazon Route 53 Launches New API Action to list Private Hosted Zones associated with your Amazon VPCs – The Route 53 database now makes working with foreign keys way easier. You used to have to write custom scripts for this.
Amazon SES can now send notifications when the delivery of an email is delayed – That notification comes via SNS, which you can configure to report to you via email. And so the circle of life continues, with you getting emails about your emails. And now I’m telling you about it via email.
Amazon WorkDocs refreshes the iOS app mobile UI to provide a new, simplified user experience – Now I’m just waiting for the slapdown court fight when Apple demands 30% of AWS revenues for allowing this app in their app store.
AWS Amplify Console now supports deploying and hosting web apps managed in monorepos – This feels like a facepalm moment for the Amplify team; post-shipping, they talk to a customer who uses repositories differently than they do.
AWS App Mesh controller for Kubernetes is now generally available – The line between “overly complicated architectural diagrams of your AWS environment” and “Kubernetes reference architecture” continues to blur.
AWS AppConfig announces the launch of hosted configurations – “Manage a configuration file yourself and supply it every time you want to update something” may be how CloudFormation, most Lambda deployment frameworks, the Route 53 API, and a few other things work, but thankfully AppConfig no longer rolls that way.
AWS DataSync can now transfer data to and from AWS Snowcone – “C’mon, we’re behind in our ‘number of feature releases’ quota for the quarter” whines the DataSync manager who’s presumably bonused on this. “Let us announce compatibility with the thing the same day as you release it. Be a pal!”
Detailed Cost Management Data is now available on AWS Console Mobile Application – Comprehensible detailed cost management data remains only available within Excel, Tableau, or third party services.
Introducing AWS CloudFormation Guard (Preview) – a new open-source CLI for infrastructure compliance – Ooh, another open source project from AWS where the only committers are AWS employees! I bet we can change that; line up to help if your idea of a good time is doing volunteer work for a trillion dollar company.
Amazon Builders Library | Automating safe, hands-off deployments – A wonderful dive into how to automate safe deployments. Unlike most papers on the subject, this paper has no vendor solution to sell you. If anything, it highlights a couple of deficiencies in the AWS managed tooling for CI/CD…
Amazon EKS Now Supports EC2 Inf1 Instances | AWS News Blog – By pairing machine learning instances with Kubernetes, AWS is addressing the customer pain of “having way too much money to spend on nonsense.”
Introducing AWS Snowcone – A Small, Lightweight, Rugged, Secure Edge Computing, Edge Storage, and Data Transfer Device | AWS News Blog – This solves the biggest problem that the AWS Snowball Edge had that wasn’t “customers looking up the product name on Urban Dictionary.” Specifically, when you wanted to hurl data through someone’s window out of rage and/or spite, you used to need a trebuchet to do it. Instead of weighing 50 pounds, the SnowCone clocks in at 4.5 pounds for 8TB of storage–or 111GB per ounce. And if you’re not calculating your storage per pound metrics, you’re probably doing it right.
New – A Shared File System for Your Lambda Functions | AWS News Blog – You can now mount the same file in Lambda, Fargate, and EC2. Let a thousand terrible patterns bloom alongside the four good ones that this empowers.
Detecting and visualizing telecom network outages from tweets with Amazon Comprehend | AWS Machine Learning Blog – This is handy; with a bit of tweaking it’ll let you detect AWS service outages long before stop.lying.cloud is updated!
Running a business is hard. Your cloud doesn’t have to be. DigitalOcean is the cloud that offers transparent, predictable pricing – even for Kubernetes clusters, which you’d have thought was impossible! You also won’t need 12 weeks of cloud school to absorb a zillion ancillary services just to be able to SSH into an instance. Is this the kind of simplicity you need out of your cloud provider? Check out DigitalOcean today.
We’ve talked about this previously a couple of times, but now that I have need for using it, it’s time once again to highlight the best way to tune your Lambda RAM allocation for your workloads.
It’s sad that awsls has to exist, but I’m glad it does. The AWS CLI maddeningly switches between “get” “describe” and “list” to enumerate resources on a service-by-service basis.
A handy guide on using AWS Lambda functions as cron jobs. It’s unclear whether “AWS Lambda” or “cron syntax” are more obscure.
… and that’s what happened Last Week in AWS.