After replacing ClusterCacheTracker with ClusterCache in #11247, there isn't any documentation about ClusterCache in the CAPI book nor in repo.

This documentation gap became evident as we got a question in cluster-api-provider-metal3 that why CAPM3 is connecting to cluster-api-provider-proxmox clusters, if both are created in same namespace, but proxmox doesn't connect to CAPM3 clusters. After a bit of digging, we came to conclusion that CAPM3 implementing CAPI ClusterCache is behind this, and since Proxmox provider doesn't implement ClusterCache, it explains why the connection doesn't happen the other way around.

I'm opening this documentation issue for the lack of ClusterCache documentation for the provider developers. Documenting this default behavior plus watch filters, and their expected uses shortly would go a long way here.

/kind documentation