Kubernetes architecture consists of master/control plane components and worker nodes, working together to manage containerized applications. Here's a brief breakdown:
Control Plane (Master Node):
- API Server: Provides communication and control via the Kubernetes API.
- Scheduler: Distributes workloads (pods) among nodes according to the availability of resources.
- Controller Manager: Manages endpoints, replicas, and other control loops.
- etcd: A cluster configuration and state key-value store.
- Cloud Controller Manager: Connects cloud-specific services to Kubernetes.
Worker Nodes:
- Kubelet: Verifies that containers are operating as intended by running on each worker node.
- Kube-Proxy: Network rules are managed by Kube-Proxy to - - facilitate pod communication.
- Container Runtime: In charge of executing containers, such as Docker and containerd.
Additional Crucial Ideas:
- Pods: The smallest deployable units are called pods, which hold one or more containers.
- Nodes: Workload-running computers, either physical or virtual.
- Services: Load balancing pods and stable endpoints for exposure.
- Persistent Storage: Volumes for managing data over restarts are known as persistent storage.
- Namespaces: Multi-tenancy logical divisions.
To efficiently manage the cluster, Kubernetes makes use of declarative configurations and intended state reconciliation.
Top comments (0)