Kubernetes 容器编排架构

Architecture

Kubernetes 容器编排架构图，展示控制平面（API Server、etcd、Scheduler、Controller Manager）、工作节点（Kubelet、容器运行时、kube-proxy、Pod）、网络层（Ingress、Network Policy、Service Mesh）和带 CSI 驱动的持久存储。该模板提供从控制平面到存储层的完整 Kubernetes 架构栈视图。对于管理 Kubernetes 集群的 DevOps 工程师和平台团队来说是基础参考。

下载 .fz 文件

完整 FlowZap 代码

ControlPlane { # Kubernetes Control Plane
n1: rectangle label:"API Server"
n2: rectangle label:"etcd (Cluster State)"
n3: rectangle label:"Scheduler"
n4: rectangle label:"Controller Manager"
n5: rectangle label:"Cloud Controller"
n1.handle(right) -> n2.handle(left) [label="Store State"]
n1.handle(bottom) -> n3.handle(top) [label="Schedule Pods"]
n3.handle(right) -> n4.handle(left) [label="Reconcile"]
n4.handle(right) -> n5.handle(left) [label="Cloud Resources"]
n3.handle(bottom) -> WorkerNode.n6.handle(top) [label="Assign Pod"]
}
WorkerNode { # Worker Node
n6: rectangle label:"Kubelet"
n7: rectangle label:"Container Runtime (containerd)"
n8: rectangle label:"kube-proxy"
n9: rectangle label:"Pod A (App Container)"
n10: rectangle label:"Pod B (App Container)"
n11: rectangle label:"Pod C (Sidecar + App)"
n6.handle(right) -> n7.handle(left) [label="Pull Image"]
n7.handle(right) -> n9.handle(left) [label="Start Container"]
n7.handle(bottom) -> n10.handle(top) [label="Start Container"]
n7.handle(bottom) -> n11.handle(top) [label="Start Container"]
n8.handle(bottom) -> Networking.n12.handle(top) [label="Service Routing"]
}
Networking { # Networking Layer
n12: rectangle label:"ClusterIP Service"
n13: rectangle label:"Ingress Controller"
n14: rectangle label:"Network Policy"
n15: rectangle label:"Service Mesh (Istio)"
n12.handle(right) -> n13.handle(left) [label="Expose"]
n13.handle(right) -> n14.handle(left) [label="Filter"]
n14.handle(right) -> n15.handle(left) [label="mTLS"]
}
Storage { # Persistent Storage
n16: rectangle label:"PersistentVolume"
n17: rectangle label:"StorageClass (SSD/HDD)"
n18: rectangle label:"CSI Driver"
n16.handle(right) -> n17.handle(left) [label="Provision"]
n17.handle(right) -> n18.handle(left) [label="Cloud Disk"]
n16.handle(top) -> WorkerNode.n9.handle(bottom) [label="Mount Volume"]
}

为什么需要这个工作流？

Managing containers across multiple hosts manually is unsustainable beyond a few services. Kubernetes automates container deployment, scaling, networking, and storage orchestration—providing a declarative platform where you describe the desired state and the system continuously reconciles to achieve it.

工作原理

Step 1: The API Server receives deployment manifests and stores desired state in etcd.
Step 2: The Scheduler assigns pods to worker nodes based on resource requirements and constraints.
Step 3: Kubelet on each worker node pulls container images and starts pods.
Step 4: kube-proxy manages network routing and service discovery within the cluster.
Step 5: Ingress controllers expose services to external traffic with TLS termination.
Step 6: PersistentVolumes provide durable storage that survives pod restarts and rescheduling.

替代方案

Docker Compose works for single-host deployments. ECS/Fargate provides managed container orchestration without Kubernetes complexity. Nomad offers a simpler alternative for some workloads. This template provides the comprehensive Kubernetes architecture reference.

Key Facts

Template Name	Kubernetes 容器编排架构
Category	Architecture
Steps	6 workflow steps
Format	FlowZap Code (.fz file)