Add scheduler plugins helm chart

[carlory]

Test:

(base) ➜  scheduler-plugins git:(install-as-a-second-scheduler) helm upgrade --install llmaz-scheduler --set scheduler.image.repository=ghcr.io/carlory/inftyai/scheduler-plugins --set scheduler.image.tag=latest  ./chart
Release "llmaz-scheduler" has been upgraded. Happy Helming!
NAME: llmaz-scheduler
LAST DEPLOYED: Thu May 29 18:11:25 2025
NAMESPACE: default
STATUS: deployed
REVISION: 9
TEST SUITE: None

(base) ➜  scheduler-plugins git:(install-as-a-second-scheduler) kubectl get po
NAME                                                            READY   STATUS    RESTARTS   AGE
llmaz-scheduler-6899f8dc7d-fljfj                                1/1     Running   0          8m15s

(base) ➜  scheduler-plugins git:(install-as-a-second-scheduler) kubectl get po -owide
NAME                                                            READY   STATUS    RESTARTS   AGE     IP            NODE                           NOMINATED NODE   READINESS GATES
llmaz-scheduler-6899f8dc7d-fljfj                                1/1     Running   0          9m15s   10.244.3.24   kind-worker                    <none>           <none>
qwen2-0--5b-demo-0                                              1/1     Running   0          5m50s   10.244.10.2   quirky-hofstadter-2901052955   <none>           <none>

(base) ➜  scheduler-plugins git:(install-as-a-second-scheduler) kubectl get openmodel qwen2-0--5b -oyaml
...
  inferenceConfig:
    flavors:
    - limits:
        cpu: 100m
        memory: 10Mi
      name: 0x
      nodeSelector:
        node.kubernetes.io/instance-type: c-0x-amd64-linux
    - limits:
        cpu: 100m
        memory: 10Mi
      name: 2x
      nodeSelector:
        node.kubernetes.io/instance-type: c-2x-amd64-linux
    - limits:
        cpu: 100m
        memory: 10Mi
      name: 1x
      nodeSelector:
        node.kubernetes.io/instance-type: c-1x-amd64-linux
...

(base) ➜  scheduler-plugins git:(install-as-a-second-scheduler) kubectl get po qwen2-0--5b-demo-0 -oya
ml
apiVersion: v1
kind: Pod
metadata:
  annotations:
    leaderworkerset.sigs.k8s.io/size: "1"
    llmaz.io/inference-service-flavors: 1x

(base) ➜  scheduler-plugins git:(install-as-a-second-scheduler) kubectl get nodeclaim
NAME            TYPE               CAPACITY   ZONE          NODE                           READY   AGE
default-tgf4f   c-1x-amd64-linux   spot       test-zone-a   quirky-hofstadter-2901052955   True    3h56m

[kerthcet]

rename to kube-scheduler, see https://hub.docker.com/repository/docker/inftyai/kube-scheduler/general

[carlory]

updated.

[kerthcet]

/lgtm
/approve
/kind feature

I'll merge first and will test to push a chart later.