Skip to content

Worker node

Provides a running environment for client applications Applications is incapsulated in pods. Pod is a smallest cheduling work unit, consist of containers. 💡 In a multi-worker Kubernetes cluster traffic with client users is not routed throw control plane, traffic is handled directly by the worker nodes.

Worker node Components

  • Container RuntimePasted image 20230929200501.png
  • Node Agent - kubeletPasted image 20230929201217.png
    • is an agent running on each node and communicates with control plane
    • interacts with the container runtime to run containers associated with the Pod,
    • monitors the health and resources of Pods containers.
    • connects to container runtime throw  Container Runtime Interface (CRI)
      • ImageService  for all the image-related operations
      • RuntimeService  for all the Pod and container-related operations
    • CRI shims
      • cri-containerd
      • CRI-O
      • dockershim and cri-dockerd
  • Proxy - kube-proxy
    • is a network agent running on each node and responsible for dynamic updates of all networking rules on the node.
    • is responsible for TCP, UDP, and SCTP stream forwarding
  • Add-ons for DNS, Dashboard user interface, cluster-level monitoring and logging.
    • Add-ons are cluster features and functionality not yet available in Kubernetes, but implemented by 3rd-party pods and services.
        • **DNS
      • Dashboard
      • Monitoring
      • Logging Usually Fluentd and Prometheus is used for logging and monitoring

References

  1. LinuxFoundationX LFS158x. Chapter 4. Kubernetes Architecture
  2. https://trainingportal.linuxfoundation.org/courses/kubernetes-for-developers-lfd259 Kubernetes Architecture