Machine Deployments

A Kubernetes-Cluster typically consists of a controlplane (running the Kubernetes API, the etcd database, and other components) and worker nodes - (virtual) servers, which run the actual workloads (applications etc.).

The controlplane of all GKS-Clusters is managed by the GKS platform itself and is not accessible for customers.

The worker nodes of a cluster need to be configured with a Machine Deployment. The customer can control:

• Node Flavors are the virtual machine types which are used to create the worker nodes. The types usually differ in their size, i.e. in the available CPU and RAM.
• Number of Nodes you can choose how many worker nodes you want to have in total.
• SSH-Keys if enabled for your cluster and if your nodes have a public IP (Floating IP), you can connect to your worker nodes with SSH, i.e. for extended debugging.
• Operating System of the Nodes although the OS of the worker-nodes should not matter for the Kubernetes workloads, it might matter for you and your debugging.

Learn More

• Sizing: Choosing Node Flavors
• Debugging: Cluster Nodes Usage Rate
• Debugging: Adding SSH Keys to Worker Nodes
• Upgrading the OS on Worker Nodes
• Cluster Nodes in multiple Availability Zones

Table of contents

• Machine Deployment
• Choosing Node Flavors/Sizes
• Cluster Nodes Usage Rate
• Adding an SSH Key to an Existing Cluster
• Updating the OS of Worker Nodes
• Multiple Availability Zones