myserviceexample.yaml

apiVersion: v1
kind: Service
metadata:
  name: ${SERVICENAME}
spec:
  ports:
    -
      port: 80 # internal port for communication between internal pods and services using the service name
      targetPort: 3000 # port on the POD
      #you can specify a nodePort if you know one does not exist already
      #nodePort: 31046 # port for ELB to load balance. since this gets created on each kubernetes minion node it must be unique cluster wide
      #to obtain the nodePort on an existing service run "kubectl describe svc SERVICE --namespace=dev|qa|prod"
  selector:
    service: ${SERVICENAME}
  type: ClusterIP
---
apiVersion: v1
kind: ReplicationController
metadata:
  # rc name matches service name for simplicity of rolling updates
  name: ${SERVICENAME}
spec:
  replicas: 2
  selector:
    service: ${SERVICENAME}
    build: build${BUILD}
  template:
    metadata:
      name: mytemplate-${SERVICENAME}
      labels:
        service: ${SERVICENAME} #must match the selector of the rc, used for linking pods to services
        build: build${BUILD}
    spec:
      containers: #many containers can be specified here that will be scheduled together onto the same host
      - name: my-container #first container
        image: ${DOCKER_REGISTRY}/${CONTAINER1}:build${BUILD} #container image location
        ports:
        - containerPort: 3000 #port running inside and outside the container on the pod IP address
        #  hostPort: 3000 #not recommended\required but shown here, can be used to expose the container port directly on the kubernetes node
        env: #environment variables provided to the container, there are also default ones created for each container that can be leveraged cluster wide.
          - name: "DB_SERVER"
            value: "mysql-service" #this can be a service endpoint that will be static, external value stored in a registry this is not required for this service and is only an example
            #kubernetes adds a cluster wide dns entry for any service created when using the skydns plugin
        livenessProbe: #health check configuration, will kill the container when this fails
          httpGet: #http shown here, you can also detail out a command\script to execute inside the container that expects a 200 status code.
            path: /
            port: 3000
          initialDelaySeconds: 120
          timeoutSeconds: 1
        readinessProbe: #health check configuration, will take the pod out of rotation when this fails
          httpGet: #http shown here, you can also detail out a command\script to execute inside the container that expects a 200 status code.
            path: /
            port: 3000
          initialDelaySeconds: 5
          timeoutSeconds: 1
#        command: # command to execute inside the container on creation
#        - "npm start"