With Platform Engineering, We Can Forget About DevOps | ✉️ #39
Hey! 👋
Last week we’ve released a podcast episode focused on platform engineering, with Sergiu Petean, Head of DevOps at Allianz Direct, as a guest. Check it out for the deep and insightful conversation on the topic. In this dispatch intro, I want to share some thoughts on why I love the term “platform engineering”.
A while ago, I wrote what DevOps is:
In this complicated situation engineers and those who just sympathize them had to start outreach activities. And how can one be successful in them without a catchword? That is how DevOps was born – a marketing term, which causes different associations in people's minds from "inner company culture" to "jack of all trades". <…> DevOps has naturally changed from "culture" and "ideology" to "profession".
“DevOps” doesn’t have a proper definition. It means so many things at the same time that it stopped meaning anything in particular at all. In comparison, take SRE - it has a pretty solid definition and it’s not that hard to describe what SRE Engineers are doing. But SRE doesn’t quite cover the same areas of responsibility that “DevOps engineers” do. That’s why I am happy to see the growing popularity of the term “platform engineering”. Finally, we have a proper term, that incorporates both engineering tasks and all the collaboration and product building work that modern DevOps teams do.
With platform engineering becoming the new default description to the profession part of DevOps, we can finally rollback the definition of DevOps to its original culture-heavy context, without mixing in all the possible technologies and skill sets. One day in the future, we should see more job positions for platform engineers and less for DevOps engineers. DevOps, like Agile, will stick to the domain of “things that ideally whole engineering department should know and practice” and not something that a special group of people is doing.
Now the final problem we have to solve is to properly define what “platform” means. But that’s a whole other discussion.
What We've Shared
On our Youtube channel the journey of cost optimization for AWS and GCP continues:
Control AWS costs with these 3 cost allocation tags. If you're looking to control your AWS costs, then these 3 cost allocation tags are a great start! Learn how to use them effectively with Terraform and unlock a new dimension inside AWS Cost Explorer.
Reduce Your Google Cloud Run Costs With These 2 Easy Tips! Cloud Run, a Google Cloud serverless tool, seems cost-fixed, but we can cut expenses. This video shows how adjusting network settings and CPU allocation can significantly lower Cloud Run costs.
DevOps Accents #31: Understanding Platform Engineering with Sergiu Petean from Allianz Direct.
And on the website it's articles galore, all kinds of themes this time around:
How to add alt text to 1000 images with GPT-4 Vision AI. We've used AI to automate alt text for images, enhancing SEO and accessibility. Using OpenAI's GPT-4 Vision with Ruby, we've streamlined image descriptions for our DevOps and Cloud content, showcasing AI's efficiency in content optimization.
Getting started with AWS Cost Optimization - 6 steps to get the cloud bill under control. From AWS's tool to analyze your costs not being enabled by default to configuring AWS Cost Anomaly Detection, there's a lot to do to get started with your AWS cost optimization. Here are the first 6 steps for you to make.
How do deploy Applications within Projects in Argo CD. In the third lesson of our Argo CD Lightning Course we are finally going to deploy the application, by first placing it in a project!
What We've Discovered
Amazon EKS introduces upgrade insights. This new feature allows to automatically get a better idea of whether new K8s version will ruin your week.
A deep dive into simplified Amazon EKS access management controls. Gone are the days of managing EKS Auth with aws-auth configMap! Starting December 2023, EKS supports new access management controls, that offer a bit more integrated and nicer way to configure auth for your Kubernetes clusters on AWS.
Train Llama2 with AWS Trainium on Amazon EKS. A practical example of using AWS Trainium. Does it conflict with other services where you can train models on AWS? Yes. But maybe you love doing it on EKS in particular, so here you go.
Streamlining and Implementing Incident Management at Dyninno. A systematic approach to handling incidents, with process diagrams included.
A Distributed Systems Reading List. Decent collections of all things distributed, including most common problems, failure models and concepts.
A random reminder
There are more places where you can listen to to our podcast. If you prefer Google Podcasts to Spotify or Apple Podcasts, here you go! Let us know, if you want to find us on some other podcast platform, if we are not there already, and we'll be there for you.
The 40th mkdev dispatch will arrive on Friday, March 15th. See you next time!