Easy Cluster mode on kubernetes

### Feature request type

Kubernetes Operator, or easy config file setup for garnet

### Is your feature request related to a problem? Please describe

Using cluster mode on kubernetes is impossible without an operator to meet the other nodes in a statefulset.
Indeed, garnet uses the same cluster model as redis where stafulset/service (with clusterIP none) dns naming pattern is unknown at the config level.

### Describe the solution you'd like

provide an operator that orchestrates cluster setup, recovery, migration based on a custom CRD that describes the cluster topology :

```yaml
apiVersion: garnet.microsoft/v1
kind: GarnetCluster
metadata:
name: garnet-cluster
spec:
masterCount: 3
replicasPerMaster: 1
storage:
volumeClaimTemplate:
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 1Gi
```

- Changing masterCount should reshard and migrate data to resized masters
- changing replicasPerMaster should replicate data

### Describe alternatives you've considered

Allow graceful recovery of the cluster when one or many nodes restart using a dns naming pattern (would be nice) using a predefined cluster DNS names lists. Example :
in garnet.conf
```json
{
"clusterMasterNames": ["garnet-0", "garnet-1", "garnet-3"],
"clusterReplicaNames" : ["garnet-4", "garnet-5", "garnet-6"]
}
```

- `clusterReplicaNames` vector should have a size that is multiple of `clustesMasterNames`
- each cluster know who he is by getting it's hostname as kubernetes would provision it whensetting a statefulset
- only static setup would be supported (no dynamic resharding, or dynamic replication)

### Additional context

_No response_

No files, tests, or code entry points are named. Start by reviewing Garnet's cluster-mode configuration alongside the proposed Kubernetes StatefulSet, headless Service DNS naming, and GarnetCluster CRD. Done would mean a defined path for cluster setup, recovery, migration, and the stated master and replica count changes.

Written by the indexing model from the issue text.