Skip to main content

Data Flow: Control-Plane HA (kube-vip)

Purpose: For platform engineers, explains how kube-vip provides a stable virtual IP for the Kubernetes API server across control-plane nodes.

Flow Summary​

Components​

ComponentNamespaceRole
kube-vipkube-systemRuns as static pod on each control-plane node, elects leader, advertises VIP
Kubernetes API Serverkube-systemServes the API on each control-plane node

Sequence​

  1. kube-vip runs as a static pod on every control-plane node.
  2. Instances participate in leader election (using Kubernetes lease or Raft).
  3. The elected leader binds the virtual IP to its node's network interface.
  4. Leader advertises the VIP via ARP (L2 mode) or BGP (L3 mode).
  5. All cluster nodes, kubectl clients, and Headlamp connect to the API server through this VIP.
  6. If the leader node fails, a new leader is elected and the VIP migrates within seconds.

Why It Matters​

Without kube-vip, clients would need to know individual control-plane node IPs. The VIP provides:

  • A single stable endpoint for the kubeconfig server field
  • Automatic failover if a control-plane node goes down
  • No external load balancer dependency for the API server