Good Morning!
Later this week it becomes February, and AWS will begin charging half a penny per hour for every public IPv4 address you have. My own bill is going to rise by nearly 10% as a result.
It’d be a lot less galling if AWS didn’t have a sad, sad level of coverage for pure IPv6 stacks among their own services. In turn that makes it feel a lot less like the aspirational thing I thought it was when I penned that blog post last summer, and a lot more like a nakedly transparent cash grab. They had six months to shore up v6 coverage–and they failed to do so.
From the Community
I recently stumbled across this article from last year, The Meaning(lessness) of Serverless. Ben Kehoe urges us not to get worked up about how "Serverless" no longer means "doesn’t scale to zero," but I get worked up about it all the same. For a shining moment, AWS was a paragon of developer friendly infrastructure that charged you nothing when you weren’t using it, but that gave way to the company absolutely tripping all over itself to get another few dollars a month from virtually everyone.
I really needed this reminder that corporations are not to be loved. Yes, it’s about Apple in this case, but it applies universally.
Martin Fowler has opined at length on the current state of Continuous Integration. This is worth consuming over a cup of coffee.
I liked this economic comparison of background jobs running on Lambda vs ECS vs ECS Fargate. It’s really hard to get realistic comparisons wherein you’re comparing apples to apples instead of orangutans, but the author did an admirable job of controlling for that.
Fathom Analytics has a post up about how they methodically reduced their AWS bill by $100,000. I like Fathom’s ethos, and their approach is directionally correct; understand the big line items in the bill, figure out what each sub-item means, and figure out if your current configuration is serving you in ideal ways. From there, work your way down the bill from largest numbers to smallest, and you’ll come out ahead by the time that you’re done.
Podcasts
Last Week In AWS: A Welcome EKS Price Hike
Screaming in the Cloud: How Snyk Gets Buy-In to Improve Security with Chen Gour Arie
Screaming in the Cloud: Working to Live Instead of Living to Work with Jeremy Tanner
Choice Cuts
AWS Billing Conductor releases account-scoped custom line items – This is really exciting for a couple dozen big enterprises, my day job, a raft of channel resellers that I generally spend approximately zero time thinking about, and absolutely nobody else. This is a confusing addition to a service that most folks don’t really need to think about, but if you’re a big company wrestling with presenting a custom view of your AWS bill to your internal teams, there’s nothing quite like Billing Conductor. It’s just got a very steep learning curve, and it sounds ridiculous if the problems it solves aren’t ones that you’re experiencing.
Introducing IPv6 instance bundles on Amazon Lightsail – This really feels like someone dropped a ball and forgot LightSail existed. The price hike for IPv4 hits Lightsail on May 1, not Feb 1 along with everything else. Oops.
Sellers can now resell third-party professional services in AWS Marketplace – AWS has just hinted at its ambition to become the eBay of Cloud Services with their new marketplace feature, but hopefully without the "executives don’t like someone who wrote a newsletter critical of them so they harassed the bejeezus out of the author until it wound up in court" part. They’re now letting sellers hock third-party professional services, because who doesn’t trust the buyer-seller relationship like a gamble in an internet auction? Hope they’re working on a rating system for "mild competence" because some of us have trust issues with Amazon trying to bootstrap its own entire shadow economy.
The Journey to IPv6 on Amazon EKS: Foundation (Part 1) – IPv6 doesn’t support dual stack deployments, so rather than a three part lecture to the rest of us on what we should be doing, how about you go fix your own nonsense so this comes off as a hell of a lot less preachy? A whole bunch of folks are seeing significant hikes to their AWS bill this week, and AWS dragging its feet on getting its own services ready for IPv6 is not only an embarrassment to the company, it’s an expensive burden to customers through no fault of their own.
Deploy a Microsoft Teams gateway for Amazon Q, your business expert – Very few people like Microsoft Teams, and absolutely nobody likes Amazon Q (that useless chatbot that interrupts you in the AWS marketing pages, the AWS console, and all throughout the documentation site like Microsoft Clippy with a traumatic brain injury), so pairing the two of them just makes sense. Calling Amazon Q an expert in any way, shape, or form is doing an incredible disservice to actual experts.
RED Camera Cloud Upload to AWS – Yeah, given that RED cameras record video at roughly a gigabyte a minute, I too would want customers to shove that data into S3 were I AWS. If that data were flowing the other way, the data transfer charges would bankrupt some small countries.
VMware’s Cloud Journey: AWS Account Management at Scale – Did someone at AWS miss how since the Broadcom acquisition VMware has just ruined life for all of their employees, partners, and customers? Okay sure, they’re great at managing AWS accounts at scale–but highlighting what’s suddenly become a company that’s so customer-hostile that it makes Oracle look like they’re customer obsessed is a really strange choice by AWS.
Tools
DNS remains a great database, but it can also be a whole mess of other ridiculous things.
… and that’s what happened Last Week in AWS.
