datastax
Option to manage CRDs separately from the main cluster object
I'm running into some problems due to the large size of the cluster CRD. I sometimes get memory errors in K9s/kubectl, and it can be a bit difficult to find the config I'm looking for. I would like to have a mode where the cluster CRD just has a flag to tell it to manage a proxy CRD. That way I can work directly with the proxy config without affecting config of the other components.
This is not doable since the internal CRDs are not meant for it.
What are the errors you're seeing? We could look at reducing the CRD size but it's really hard now.
One thing I've noticed with kubectl is that using --server-side helps
I get errors when using kubectl edit to edit the PulsarCluster resource directly. I realize this is probably a big request, so I guess this one can be pushed back, it's more just a convenience.
E363: Pattern uses more memory than 'maxmempattern'
As far as implementation, I was thinking you could just set a flag in the PulsarCluster to say managedIndependently or something for the proxy for example, and then the operator just picks up config directly from the Proxy custom resource.