ACM started using ClusterPools as a poor man's templating engine. They noticed they could specify the necessary configuration just once in the pool spec, and then by creating ClusterClaims, stamp out identical ClusterDeployments on demand.

One problem with this approach – with ClusterPools in general, really – is that every time ClusterPool needs to support a new ClusterDeployment field, we have to add it manually to the ClusterPool spec.

What we would like to be able to do is coverge on a more unified solution where
a) we have native support for templating ClusterDeployments; and
b) ClusterPools are used for their originally-intended purpose: to maintain a pool of ready clusters.

More details to be outlined in a formal design document.