This week's articles
Multi cluster monitoring with Thanos
In this post, Banzai Cloud explains how they settled on Thanos as a standardised solution that would allow them to federate metrics, and collect them into a single place for long term storage, querying and analysis.
This GitHub repo contains a reference implementation of a fully productionalized Kubernetes setup, based on immutable infrastructure and Gitops Flow methodologies. Everything starts out in Git as either code or configuration. Items are manipulated to the desired state and that is applied onto the infrastructure.
kube-resource-report is a handy script that generates an HTML report of cluster and pod resource requests (CPU and memory) vs. usage (collected via Metrics API/Heapster). If you are curious, you can check out the sample HTML report
Little utility that provides an easy way to give all your EC2 instances SSH host certificates: run it in your EC2 userdata script, and it will use AWS KMS to sign the instance's SSH host key. You can then add a @ cert-authority line to your ~/.ssh/known_hosts and never be prompted about an unknown host again.
conman - [the] container manager: inception
The first article in a series covering the implementation of a high-level container runtime. If you are wondering what is a container manager, some prominent examples would be containerd, cri-o, dockerd, and podman.