You are browsing unreleased documentation. See the latest documentation here.

Deploying Sidecars

Kong Gateway Operator uses PodTemplateSpec to customize deployments.

Here is an example that deploys a Vector sidecar alongside the proxy containers.

Configure vector.dev

apiVersion: v1
kind: ConfigMap
metadata:
  name: sidecar-vector-config
data:
  vector.toml: |
    [sources.proxy_access_log_source]
    type = "file"
    include = [ "/etc/kong/log/proxy_access.log" ]
    [sinks.proxy_access_log_sink]
    type = "console"
    inputs = [ "proxy_access_log_source" ]
    encoding.codec = "json"

Configure PodTemplateSpec

Using DataPlane

This method is only available when running in hybrid mode

The DataPlane resource uses the Kubernetes PodTemplateSpec to define how the Pods should run.

apiVersion: gateway-operator.konghq.com/v1beta1
kind: DataPlane
metadata:
  name: dataplane-example
  namespace: kong
spec:
  deployment:
    podTemplateSpec:
      metadata:
        labels:
          dataplane-pod-label: example
        annotations:
          dataplane-pod-annotation: example
      spec:
        volumes:
        - name: cluster-certificate
        - name: sidecar-vector-config-volume
          configMap:
            name: sidecar-vector-config
        - name: proxy-logs
          emptyDir:
            sizeLimit: 128Mi
        containers:
        - name: sidecar
          image: timberio/vector:0.31.0-debian
          volumeMounts:
          - name: sidecar-vector-config-volume
            mountPath: "/etc/vector"
          - name: proxy-logs
            mountPath: "/etc/kong/log/"
          readinessProbe:
            initialDelaySeconds: 1
            periodSeconds: 1
        - name: proxy
          image: kong/kong-gateway:3.8.0.0
          volumeMounts:
          - name: proxy-logs
            mountPath: "/etc/kong/log/"
          env:
          - name: KONG_LOG_LEVEL
            value: debug
          - name: KONG_PROXY_ACCESS_LOG
            value: "/etc/kong/log/proxy_access.log"
          resources:
            requests:
              memory: 64Mi
              cpu: 250m
            limits:
              memory: 1024Mi
              cpu: 1000m
          readinessProbe:
            initialDelaySeconds: 1
            periodSeconds: 1

Using GatewayConfiguration

This method is only available when running in DB-less mode

The GatewayConfiguration resource is a Kong-specific API which allows you to set both controlPlaneOptions and dataPlaneOptions.

You can customize both the container image and version.

1. Define the image in the GatewayConfiguration.

   kind: GatewayConfiguration
   apiVersion: gateway-operator.konghq.com/v1beta1
   metadata:
     name: kong
     namespace: default
   spec:
     dataPlaneOptions:
       deployment:
         podTemplateSpec:
           metadata:
             labels:
               dataplane-pod-label: example
             annotations:
               dataplane-pod-annotation: example
           spec:
             volumes:
             - name: cluster-certificate
             - name: sidecar-vector-config-volume
               configMap:
                 name: sidecar-vector-config
             - name: proxy-logs
               emptyDir:
                 sizeLimit: 128Mi
             containers:
             - name: sidecar
               image: timberio/vector:0.31.0-debian
               volumeMounts:
               - name: sidecar-vector-config-volume
                 mountPath: "/etc/vector"
               - name: proxy-logs
                 mountPath: "/etc/kong/log/"
               readinessProbe:
                 initialDelaySeconds: 1
                 periodSeconds: 1
             - name: proxy
               image: kong/kong-gateway:3.8.0.0
               volumeMounts:
               - name: proxy-logs
                 mountPath: "/etc/kong/log/"
               env:
               - name: KONG_LOG_LEVEL
                 value: debug
               - name: KONG_PROXY_ACCESS_LOG
                 value: "/etc/kong/log/proxy_access.log"
               resources:
                 requests:
                   memory: 64Mi
                   cpu: 250m
                 limits:
                   memory: 1024Mi
                   cpu: 1000m
               readinessProbe:
                 initialDelaySeconds: 1
                 periodSeconds: 1
         
   

2. Reference this configuration in the GatewayClass resource for the deployment.

   kind: GatewayClass
   apiVersion: gateway.networking.k8s.io/v1
   metadata:
     name: kong
   spec:
     controllerName: konghq.com/gateway-operator
     parametersRef:
       group: gateway-operator.konghq.com
       kind: GatewayConfiguration
       name: kong
       namespace: default
   

3. Use the GatewayClass in your Gateway.

   kind: Gateway
   apiVersion: gateway.networking.k8s.io/v1
   metadata:
     name: kong
     namespace: default
   spec:
     gatewayClassName: kong
     listeners:
     - name: http
       protocol: HTTP
       port: 80