kong

APIClarity Kong plugin

Prerequisite

• APIClarity backend is running.
• Kong gateway is running in your K8s cluster, and has an Ingress gateway resource.

If you just want to try it out with a demo application, and you don't have kong installed, you can follow these few short steps that will help you to quickly setup a running environment:

1. Deploy sock-shop app:

      kubectl create namespace sock-shop
   
      kubectl apply -f https://raw.githubusercontent.com/microservices-demo/microservices-demo/master/deploy/kubernetes/complete-demo.yaml
   

2. Deploy Kong:

   • Using helm3:

      helm repo add kong https://charts.konghq.com
      helm repo update
      helm -n kong install kong/kong --create-namespace --generate-name --set ingressController.installCRDs=false
   

   • Using kubectl:

      kubectl apply -f https://bit.ly/kong-ingress-dbless
   

3. Wait for pod to be ready:

      watch kubectl get pods -n kong
   

4. Configure Ingress:

      kubectl apply -f - <<EOF
      apiVersion: networking.k8s.io/v1
      kind: Ingress
      metadata:
        name: catalogue
        namespace: sock-shop
      spec:
        ingressClassName: kong
        rules:
        - http:
            paths:
            - path: /
              pathType: Prefix
              backend:
                service:
                  name: catalogue
                  port:
                    number: 80
      EOF
   

5. Deploy APIClarity Kong Plugin:

      cd ~/go/src/github.com/openclarity/apiclarity/plugins/gateway/kong
   
      KONG_GATEWAY_DEPLOYMENT_NAME=ingress-kong \
      KONG_GATEWAY_DEPLOYMENT_NAMESPACE=kong \
      KONG_GATEWAY_INGRESS_NAME=catalogue \
      KONG_GATEWAY_INGRESS_NAMESPACE=sock-shop \
      UPSTREAM_TELEMETRY_HOST_NAME=apiclarity-apiclarity.apiclarity:9000 \
      TRACE_SOURCE_TOKEN=<use one created with external trace_source>
      deploy/deploy.sh
   

   • Note: If you installed Kong using helm, the deployment name might be different. Please change the KONG_GATEWAY_DEPLOYMENT_NAME env var accordingly.

6. Get LoadBalancer Address:

      export KONG_ADDRESS=$(kubectl get -o jsonpath="{.status.loadBalancer.ingress[0].ip}" service -n kong kong-proxy)
   

   Or on EKS get hostname:

      export KONG_ADDRESS=$(kubectl get -o jsonpath="{.status.loadBalancer.ingress[0].hostname}" service -n kong kong-proxy)
   

   • Note:
     • If you installed Kong using helm, the service name might be different.

7. Run Traffic:

      curl -H 'content-type: application/json' -H 'accept: application/json;charset=UTF-8' $KONG_ADDRESS/catalogue
      curl -H 'content-type: application/json' -H 'accept: application/json;charset=UTF-8' $KONG_ADDRESS/catalogue/size
      curl -H 'content-type: application/json' -H 'accept: application/json;charset=UTF-8' $KONG_ADDRESS/tags
   

8. Cleanup:

   1. Delete kong installation:

         helm -n kong uninstall kong 
      

      Or if not installed with helm:

          kubectl delete -f https://bit.ly/kong-ingress-dbless 
      

   2. Delete sock-shop:

          kubectl delete ns sock-shop 
      

Installation using a pre-built image

1. Choose one of the following installation techniques

   1. Script installation

      • Run the following script to add the plugin to your Kong deployment and Ingress configuration.
      • Please set all env variables accordingly:

      KONG_PROXY_CONTAINER_NAME=<name> \
      KONG_GATEWAY_DEPLOYMENT_NAME=<name> \
      KONG_GATEWAY_DEPLOYMENT_NAMESPACE=<namespace> \
      KONG_GATEWAY_INGRESS_NAME=<name> \
      KONG_GATEWAY_INGRESS_NAMESPACE=<namespace> \
      UPSTREAM_TELEMETRY_HOST_NAME=<telemetry service address> ./deploy/deploy.sh
      

   2. Helm installation

      • Save APIClarity default chart values

      helm show values apiclarity/apiclarity > values.yaml
      

      • Update the values in trafficSource.kong
      • Deploy or Upgrade APIClarity

      helm upgrade --values values.yaml --create-namespace apiclarity apiclarity/apiclarity -n apiclarity --install
      

      • A post install job will execute the installation script in your cluster

Building from source

If you would like to customize the plugin, you should run this from the plugins directory:

make docker-kong

You can then push the plugin image to any public registry, and update the helm values to use that image.

Preserving Client IP Address

Kong is usually deployed behind a Load Balancer (using a Kubernetes Service of type LoadBalancer). This can result in a loss of actual Client IP address, as Kong will get the IP address of the Load Balancer as the Client IP address.

This guide lays out different methods of solving this problem.

Instructions to uninstall the Kong plugin

Run the following script to remove the plugin from your Kong deployment. The script will rollback to Kong deployment previous pre-plugin state and remove all the supporting resources created during the deployment.

Please set all env variables accordingly:

     KONG_PROXY_CONTAINER_NAME=<name> \
     KONG_GATEWAY_DEPLOYMENT_NAME=<name> \
     KONG_GATEWAY_DEPLOYMENT_NAMESPACE=<namespace> \
     KONG_GATEWAY_INGRESS_NAME=<name> \
     KONG_GATEWAY_INGRESS_NAMESPACE=<namespace> ./deploy/undeploy.sh