Amazon announced its fourth quarter and FY 2022 results last week; Tim Bray has an analysis that’s absolutely worth reading. It’s long past time to spin off AWS; it’s ridiculous and horrible for basically every other company out there that the entire "store that everyone shops at" can lose money indefinitely because the cloud computing business props it up.
From the Community
One view to see them all! Kentik provides Cloud and NetOps teams with complete visibility into hybrid and multi-cloud networks. Ensure an amazing customer experience, reduce cloud and network costs, and optimize performance at scale — from internet to data center to container to cloud. Learn how you can get control of complex cloud networks at www.kentik.com.
I’ve done split horizon DNS before. It was annoying. Fortunately it’s now apparently a lot easier via the CDK.
AWS has pushed back their IAM billing changes until next month, and they’re Googling DeepLens. For those who don’t remember, the DeepLens was a ML powered camera that was replaced by the DeepRacer’s killer feature of "being able to follow you into the bathroom."
Lambda 3.10 is still not available as a managed Lambda runtime, and I’m sick of it. I’ve now had to actively refactor projects in Python 3.10 to avoid having to stuff the thing into Docker as my own custom runtime. Let me be very clear here: if I have to manage Docker images for serverless functions, I’m going to go with Google Cloud Run instead; the experience is better in every respect that I care about.
Jeff Barr is at it again, printing 3D tiles of AWS service logos. AWS Mahjong is likely launching soon.
Hubspot has a story of how they saved millions of dollars simply by optimizing their logging. This is a pretty common story at scale…
Business Insider reports from behind a paywall that Amazon Will Only Hire Students and Recent Graduates for Entry Level Software Engineering Positions. Uh… this is apparently Amazon’s expansion into the exciting new market of "trying to get sued for age discrimination" I guess? I mean, if it worked for IBM…
An AWS-affiliated company has filed paperwork that indicates they’ll be building two new massive data center campuses in Ohio. You never hear about an AWS region ensmallening, I’ll point out.
After a month off, I’m back on my blog post horse, starting with S3 Encryption at Rest Does NOT Solve for Bucket Negligence.
Last Week In AWS: Azure Improves Slowly
Last Week In AWS: S3 Encryption at Rest Does NOT Solve for Bucket Negligence
Last Week In AWS: Timecode Burn-In, Employee Burn-Out
Screaming in the Cloud: The Art of Effective Incident Response with Emily Ruppe
Screaming in the Cloud: The Evolution of DevRel with Jeremy Meiss
Amazon CloudWatch now simplifies metric extraction from structured logs – CloudWatch of course is a service that focuses on dollar extraction from customer wallets. I want to like it more than I do, but for that to happen it needs to become a service that’s better than it is.
Amazon MemoryDB for Redis Announces 99.99% Availability Service Level Agreement – SLAs are always interesting to me; not because of the credits you get back in event of outage (which are generally piddly compared to the damage the outage causes) but rather due to the statement of confidence in a service’s reliability.
AWS CloudTrail Lake now supports ingestion of activity events from non-AWS sources – I really like CloudTrail Lake; now that it’s growing beyond just AWS audit logs, I want to see more tooling start to integrate with it. It can start by giving us an option for the (free) management trail in all accounts having an option to redirect to this instead of its usual place.
AWS announces access of Simple Monthly Calculator estimates in the AWS Pricing Calculator – I’m very sad about the new calculator, and this feels like the old one (that has its flaws but is better suited for how humans actually use calculators like this) is going to be whacked soon.
Amazon increases NAT Gateway’s capacity to support concurrent connections to a unique destination – Yeah, at 4.5¢ and up per GB of traffic that passes through this thing, I’d want it to have higher capacity limits too.
Amazon EMR launches support for Amazon EC2 C7g (Graviton3) instances to improve cost performance for Spark workloads by 7–13% – This may be the first announcement I’ve seen about the Graviton3 chips matriculating outside of the EC2 offerings; I’ve been using one for a year or so myself and am pretty happy with it.
Analyze Amazon S3 storage costs using AWS Cost and Usage Reports, Amazon S3 Inventory, and Amazon Athena – Yes, let’s all pretend that Cost Explorer and S3 Storage Lens don’t exist. Bah. Storage Lens is great; it obviated the need for some of my own custom tooling for cost analysis.
AWS shows why physical stores matter more than ever at NRF 2023 – Amazon has shuttered its own physical stores, but don’t let a little thing like that stop you from speaking out of both sides of the corporate mouth. If customers believe that physical store are important, sure; we can say that too!
POV: You’re the little piggy who built a wolf-resistant house, and Piggy Town wants a neighborhood with your design. Can you scale quickly? Now imagine your high-transactional app is poised for massive growth. And you’re hit with a big bad traffic spike. Are you prepared?
O’Reilly’s Foundations of Scalable Systems features tips for designing scalable solutions, including replication, state management, load balancing, and caching. Get your three free chapters from Cockroach Labs — and avoid the wolves.
Depending where an arbitrary user is in the world, you can now find out what region is going to be the most responsive. This is cool.
Ruby on Jets is, of all things, a serverless framework for Ruby on Rails applications.
… and that’s what happened Last Week in AWS.