100 clusters and beyond
Posted on January 1, 2022 • 1 minutes • 192 words
Congratulations! You’re using Kubernetes. You’re company is all in and you’re no longer a systems administrator—you’re a cluster operator. You’ve set up automation to create your first handful of clusters, but it’s not enough. The infrastructure is out of hand with versions and maintenance you can’t keep up. No amount of bash is going to scale to the amount of clusters you need. So what other options are there? How can you manage 100, 500, or 1000 clusters? This talk will show patterns for deploying and managing hundreds and thousands of clusters. Terraform, CDK, Crossplane, GitOps, and more will all be discussed to show where they shine and where they might have limitations.
Benefits to the ecosystem
Many companies are currently struggling with the amount of Kubernetes clusters they’re managing. They are applying the same principles that were useful for a dozen clusters to hundreds and they can’t keep up with releases or patching. This talk will show the common patterns at different scales and when different tools might need to be used to help reach the next scale.
Open source projects
https://github.com/fluxcd/flux https://github.com/crossplane/crossplane https://github.com/aws-controllers-k8s/community https://github.com/kubernetes-sigs/cluster-api