SRE Newsletter

This week we saw some great opinion pieces on RCAs and using versioning to encourage stable software. We also got some good technical content on containers.

Next week is Amazon's re:Invent conference. We're seeing a couple sneak previews this week and should get some really interesting features in the next SRE Newsletter.

Learning Containers From The Bottom Up: // iximiuz.com
Want to learn how containers REALLY work? Ivan Velichko shares another great post where he gives a learning plan for how he learned containers in depth. The article is a reasonable length, but links to a lot of other resources so that you can get a better understanding of what's happening under the hood.
Root Cause Is for Plants, Not Software: // verica.io
This was a good one... To quote a quote from the article - "Instead of increasing safety, post-accident remedies usually increase the coupling and complexity of the system."
Stable Software Release System: // blog.kronis.dev
What does going from version 7.1.101 to 8.9.29 mean to you? Does version 2021-stable-29 make more sense?
Ask HN: Most Interesting, Mildly Impractical, Well-Written Books on Software?: // news.ycombinator.com
Everyone loves book lists, and it is close to that gift buying time of year... This list, however, intentionally skips the normal, "must read" books and gives you impractical ones instead.
DevOps in Academic Research: // mattsegal.dev
Matthew Segal documents his journey using DevOps to help generate statistical models in academic research more quickly. Overall, just a good story about a DevOps transformation that's not much different than any other environment.
Introducing IPv6-only Subnets and EC2 Instances: // aws.amazon.com
Will we ever go all IPv6? AWS is helping us get one step closer.
Amazon EKS Connector, Now Generally Available: // aws.amazon.com
Amazon annouces GA for EKS Connector. You can now extend the EKS console to visualize your Kubernetes clusters outside of AWS.
I Test in Prod: // increment.com
This article feels like it tries way too hard, but if I got the gist: the real world is not reproducible and we should embrace it. Since you can't test for everything, building for observability is key. Since perfection is not possible, error budgets are necessary.
Microservices Architecture on Google Cloud: // cloud.google.com
This post is super basic and kinda garbage, but there's a good little infographic.