Online Services NodeProblemDetector Dynamic Scheduler DynamicQuotaManager MultiClusterManager De-Scheduler Rosource Manage & Schedule Ceres Job Queue Manager Spark-Operator OfflineJobs Scheduler Kube-apiserver Pod Pod NodeResourceUsage Annotator Patch node history reource usage annotation NPD & de-scheduler Ø NOde Problem Detector (NPD) • FDPressure • CustomPIDPressure • ThreadPressure • KubeletProblem • ContainerdProblem • KernelDeadlock • … Ø Distributed Ping Test • Self-agent • Self-holder de-scheduler npd self-holder Pod Kernel Kubelet Dockerd Node Kube-ApiServer List&Watch Events

from local-in to PREROUTING • The challenges • Skb’s pointer to route is NULL during PREROUTING • No de-fragment is done during PREROUTING IPVS bypass conntrack (con.) • Egress • Original way • Nf local-out

in association with the U.S. Department of Defense to drive open source innovation into strategic de- fense initiatives. The company is delivering technology solutions that give warfight- ers the information

http://localhost:8001/api/v1/proxy/namespaces/kube-system/services/heapster/api/v1/model/namespaces/de fault/pods/cpu-demo/metrics/cpu/usage_rate 即可看到监控信息。 在本例中，尽管容器启动时，尝试使⽤2个CPU单位，但由于配置了只允许使⽤1个CPU单位，因此，最终最多只能使 http://localhost:8001/api/v1/proxy/namespaces/kube-system/services/heapster/api/v1/model/namespaces/de fault/pods/memory-demo/metrics/memory/usage 可看到如下结果： { "timestamp": "2017-06-20T18:54:00Z",

Hoenig Selvakumar Natesan Shangqi Liu Sofia Tania Vanya Seth Bharani Subramaniam Fausto de la Torre Mike Mason Maya Ormaza Birgitta Böckeler Hao Xu Neal Ford Pawan Shah Brandon Byars Ian