Dynamic Pod Resource Boundary Adjustment in Web Scale Clusters
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don’t miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
Dynamic Pod Resource Boundary Adjustment in Web Scale Clusters - Cheng Wang & Xiaoyu Zhang, Alibaba Have you ever confused about how to set perfect resource limit for Pod? How do you balance resource efficiency with application’s SLO? In this talk, we will share practices and lessons learned from adjusting Pod resource limits dynamically for Web-scale clusters at Alibaba Group by co-locating Pods with different QoS classes on the same node and adjusting Pod resource limits dynamically (especially during resource contention). After applying this effort in production clusters, we were able to improve the cluster resource usage by 14%~30%, tail latency (95 percentile) by 76%~87%, and TPS (transactions per second) by 107%~163%, respectively. The audiences would benefit from experience of improving the resource utilization and application performance for their own clusters, with Kubernetes native approaches.