see this example pod in kind. in this you can also add command and env variables. it will not restart the container itself. not complete but you will get the idea.

apiVersion: v1
kind: Pod
metadata:
  name: pod-elasticsearch-1
  namespace: common-data
  labels:
    elasticsearch-name: pod-elasticsearch-1
    node: elasticsearch
    type: elasticsearch-logs
spec:
  containers:
    - resources:
        limits:
          memory: "24Gi"
          cpu: "2"
        requests:
          memory: "16Gi"
          cpu: "1"
      ulimits:
        memlock:
          soft: -1
          hard: -1
        nofile:
          soft: 65536
          hard: 65536
      image: docker.elastic.co/elasticsearch/elasticsearch:5.5.2
      name: elasticsearch-custom
      command: [ "elasticsearch"]
      args: ["-E" , "cluster.name=es-ocp-app-cluster",
                "-E" , "node.name=master-es-node-1",
                "-E" , "node.master=true",
                "-E" , "node.data=true",
                "-E" , "network.publish_host= svc-elasticsearch-1 ",
                #"-E" , "network.host=_site_",
                "-E" , "http.host=0.0.0.0",

Create a deployment config for these requirement. In which you can declare environment variables, commands, declare secrets etc. It will be easy to maintain too.