Manage Cluster Nodes

⚠️ Gitpod Self-hosted has been replaced with Gitpod Dedicated, a self-hosted, single-tenant managed service that runs in your private cloud account but is managed by us.
Try out Gitpod Dedicated.

Sometimes nodes become unhealthy, or you need to prevent the autoscaler from removing the node from your cluster.

Avoiding Node Scale-down

If you wish to cordon a node with terminating workspaces, or, keep a node so you have time to manually backup user data:

language: 

bash

copy code

copied successfully

# reference: https://github.com/kubernetes/autoscaler/blob/master/cluster-autoscaler/FAQ.md#how-can-i-prevent-cluster-autoscaler-from-scaling-down-a-particular-node
$ kubectl annotate node <nodename> cluster-autoscaler.kubernetes.io/scale-down-disabled=true

Handling Unhealthy Nodes

Prevent new workspaces from being scheduled to a node if they become unhealthy:

language: 

bash

copy code

copied successfully

$ kubectl cordon <nodename>