Feature Overview

Allow the management of individual and specific nodes in NodePools.

Goals

HCP doesn't expose individual nodes for their management to solve this directly, and instead a series of steps are required, such as mark nodes as unhealthy or use annotations.

The goal is to have a method in HyperShift that exposes nodes individually for their management.

Use Cases

There are use cases where a specific node has to be removed from the NodePool in an automated way, declaratively from GitOps, for example. 

An OpenShift admin wants to install and manage HCP clusters using the SiteConfig Operator (ClusterInstance API). For this a template is created for the cluster and for each host. When a host is added a new set of resources for that host is created. When a host is removed the resources that were previously created for that host are removed. HCP doesn't fit in this model because it can't map a set of well-defined templates to a specific host/node.