2000 Nodes and Beyond: How We Scaled Kubernetes to 60,000-Container Clusters and Where We're Going Next
![2000 Nodes and Beyond: How We Scaled Kubernetes to 60,000-Container Clusters and Where We're Going Next](/2016/images/all/lf_huffc03acb4b89c823f315cae16e4b2e6b_29065_900x500_fit_q75_box.jpg)
Kubernetes supports 2000-Node clusters - that statement was a part of the Kubernetes 1.3 release announcement. Thats great, but what exactly does it mean? During this talk I will explain what work …
Talk Title | 2000 Nodes and Beyond: How We Scaled Kubernetes to 60,000-Container Clusters and Where We're Going Next |
Speakers | Marek Grabowski (Site Reliability Engineer, Google) |
Conference | KubeCon + CloudNativeCon North America |
Conf Tag | |
Location | Seattle, WA, United States |
Date | Nov 7- 9, 2016 |
URL | Talk Page |
Slides | Talk Slides |
Video | |
“Kubernetes supports 2000-Node clusters” - that statement was a part of the Kubernetes 1.3 release announcement. That’s great, but what exactly does it mean? During this talk I will explain what work we did to scale Kubernetes from 100 nodes in the 1.0 release to 2000 nodes, and describe the low-cost measurement infrastructure we’ve built to continuously measure Kubernetes performance at different cluster sizes. I’ll also talk a bit about our current work and plans for the future.