remove kube-rbac-proxy container from deployment

[leelavg]

csi-operator is not exporting any metrics exclusively and orchestrated csi already exports metrics on it's own which isn't tunneled through the operator.

kube-rbac-proxy prominently used to authenticate api access to metrics endpoint available on operator (manager) which isn't mandatory for us and this commit removes the same.

note that controller-runtime itself exports some metrics and this commit doesn't bind any port for controller-runtime which disables inbuilt metrics being available.

ref: https://book.kubebuilder.io/reference/metrics-reference

fixes: #119

[leelavg]

Testing:

> k logs ceph-csi-operator-controller-manager-7d64d95794-ww4jb --since 1m | tail -n3
2024-10-15T05:09:18Z    INFO    Starting workers        {"controller": "clientprofile", "controllerGroup": "csi.ceph.io", "controllerKind": "ClientProfile", "worker count": 1}
2024-10-15T05:09:18Z    INFO    Starting workers        {"controller": "clientprofilemapping", "controllerGroup": "csi.ceph.io", "controllerKind": "ClientProfileMapping", "worker count": 1}
2024-10-15T05:09:18Z    INFO    Starting workers        {"controller": "driver", "controllerGroup": "csi.ceph.io", "controllerKind": "Driver", "worker count": 1}

[nb-ohad]

@leelavg @Madhu-1 I would prefer if we would not just remove the code and configuration but make it optional. I can predict that very soon we will need to reintroduce this to support custom metrics.

[nb-ohad]

/hold

[Madhu-1]

I can predict that very soon we will need to reintroduce this to support custom metrics.

@nb-ohad can you please provide details about it?