for fast scaling Low price tag on operations No bottlenecks throughout the season 17% Record breaking Second screen interaction 62% Conclusion ● Documentation and procedures ● Prometheus ● Scaling

VWA can quickly adjust offline containers‘ cgroup to ensure the SLA of online tasks without breaking offline tasks. Key Features: Ø In-Place Update without restarting container Ø High-performance

• 错误定位 配置版本，发布，回滚，可以更加方便 微服务的开发阶段 统一采用gRPC协议和protobuf编解码 CI check 阶段 • 主要做 pb 的 format、lint、breaking 检查。 CI build 阶段 • 会基于 pb 的注释自动产生文档，并推送至内部的微服务管理系统接口平台中 • 会生成 Go/PHP/Node/Java 桩代码和错误码，推送到指定的仓库

kubectl already in production; TEE-based secure kubectl completed PoC • API server and the rest changes for the TEE-based - PoC stage • Where we go? • To keep the production practice • To explore more intersections between Cloud-Native and Confidential Computing • To submit KEPs to upstream for related changes Special Thanks We’d like to thank the contributors to this work: Yinyong Zheng, Wei Zhang, Yong

kubernetes and all that stuff • Pachyderm can version datasets and trigger new trainings when changes occur • Distributed File Systems • NFS • HDFS • … Classic DevOps solutions: • Containers • CI/CD

kubectl 功能与kubectl保持一致 3. Multi-cluster management 多集群管理 4. Improved security 提高安全性 Top requested changes 1. Third-party plugins or integrations 集成第三方插件 Which third-party plugins or integrations would

app=MyApp Replicas: 2 Service <10.0.0.11>:<9376> Label: app=MyApp Endpoints: track backend pod changes <172.17.10.1>:<80> <172.17.10.2>:<80> ... Cluster DNS record: pod1.clusterdomain pod2.clusterdomain

users looking to try out Kubernetes or develop with it day-to-day. 7 Out of tree components, what changes for me? In-tree vSphere Cloud Provider vSphere CSI Out-of-tree vSphere Cloud Provider

Prometheus + Grafana + EFK ● Ingest metrics and logs into Stackdriver without any instrumentation changes ● Aggregate logs from many clusters -- whether GKE or GKE On-Prem Logging and Monitoring Cloud

c Controller • The heart of Kubernetes orchestrator • drives the cluster state based on the changes to the API objects for { desired := getDesiredState() current := getCurrentState() makeChanges(desired