State of the UI: Leveraging Kubernetes Dashboard and Shaping its Future Dan Romlein UX Designer, Google Spencer Sugarman UX Researcher, Google Talk Goals 1. What Dashboard is and why you should

LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN # 本地环回口 inet 127.0.0.1/8 scope host lo 2: ens33: mtu 1500 qdisc pfifo_fast state UP #k8s服务器本身用于通信的网口 inet 1500 qdisc noqueue state DOWN #docker默认的容器网络 inet 172.17.0.1/16 brd 172.17.255.255 #如果不使用docker，则没有此接 口 5: kube-ipvs0: mtu 1500 qdisc noop state DOWN #Service forever preferred_l� forever 6: flannel.1: mtu 1450 qdisc noqueue state UNKNOWN #vxlan设备 inet 10.244.0.0/32 brd 10.244.0.0 scope global flannel.1 7: cni0:

0 码力 | 126 页 | 4.33 MB | 1 年前

3

Cluster = “Desired State Management” – Kubernetes Cluster Services (w/API) • Node = Container Host w/agent called “Kubelet” • Application Deployment File = Configuration File of desired state • Container all state known about cluster • Kubernetes Front-end Control Plane • Provides RESTful interface • Returns state objects as JSON • Provides core control loops for platform • Watches shared state through Clusters Desired state of Application The difference between PKS and Kubernetes Open Source Project – Google/Pivotal/VMware 21 Container scheduling, scale, resiliency, and Day 2 Desired state of Kubernetes

resource to desired config Operator: Advantages • Declarative system • Manage resource to final state continually • kube-apiserver oriented programming • CustomResourceDefinition (CRD) • Built on Kubernetes scale & offline Nodes through posting Machine CRs. • Node-Operator: difference Machine and Node state, manage Node softwares and configure files. • Machine: the instance of Machine CRD with node information, which represent a node desired in the Kubernetes. • NPD(Node Problem Detector): post Node state to kube- apiserver. Node-Operator: Scale Nodes Node-Operator Node-Operator: Upgrade Nodes Node-Operator

���������� �� ���� ������ ��������� Kubernetes Focus on long running service. Driving current state towards desired state with control loops YARN Focus on scheduling jobs ���������� ���� Kubernetes Container Only application-level metadata persisted. Container data is not persisted. Recover from in-memory state from peers ��������� ������� kubernetes CRI compatible. Docker etc. YARN Docker + TAR ball ����������

virtualservice: proxy cluster destinationrule:
 proxy-helloworld 10 90 Think in Cloud . 北北京 监控系统 Kube-State-Metrics Kubernetes APIServer cAdvisor Node-Exporter KUN-Agent Grafana Blackbox Exporter Prometheus 微信/邮件 外部探测 Probe Exporter AZ 2 (Local File) Alert Manager gossip Kubernetes 互相监控 AZ 1 Kube-State-Metrics Kubernetes APIServer cAdvisor Node-Exporter KUN-Agent Alert Manager Prometheus (Local 采集机器器 Metrics； D. 每个 Node 上以 DaemonSet 的形式部署⾃自研 KUN-Agent 采集⽹网络、⽂文件读写等 Metrics； E. 运⾏行行 Kube-State-Metrics，从它拉取 Job, Deployment 等资源的 Metrics； F. 通过 Blackbox Exporter，实现服务的⿊黑盒主动拨测； G. ⾃自研 Probe

or configures off- cluster resources • Operator waits for managed resources to reach a healthy state • Operator conveys readiness of application or managed resources to the user leveraging the status Operand metrics/alerts/logs • Operator can prevent the Operand from transitioning into an unhealthy state based on Operand metrics Auto-tuning • Operator is able to automatically tune the Operand to a certain

（创建Deployment从⽽升级ReplicaSet）。 ReplicaSet在后台创建 Pod。检查升级的状态，看是否成功。 通过更新Deployment的PodTemplateSpec来 Declare the new state of the Pods （声明Pod的新状态）。 这样，会 创建⼀个新的ReplicaSet，并且Deployment会以受控的速率，将Pod从旧ReplicaSet移到新的ReplicaSet。每个新 memory: 50Mi State: Running Started: Tue, 07 Jul 2015 12:54:41 -0700 Last Termination State: Terminated Exit Code:

Requests Application architecture Application architecture PHP-based Admin application Shared state in Redis and SQL database Scalable userfacing API in GoLang Queuing system Application technologies

networkCIDR: “172.20.0.0/16” masterPublicName: “api.nivenly.com” So how does it work ? ? ? Kops 1.4 State Store Unmarshal Tasks Map Compare Reconcile Kops 1.4 tasks := []*Task We have to track our work