Health Checks

Setup active and passive health checking using the Kong Ingress Controller. This configuration allows Kong to automatically short-circuit requests to specific Pods that are mis-behaving in your Kubernetes Cluster.

When you setup a passive health check for a service that runs in a cluster and if the Pod that runs the service reports an error, Kong returns a 503, indicating that the service is unavailable. Kong does not proxy anymore requests. There are a few options that you can consider:

• Delete the current Pod; Kong then sends proxy requests to the new Pod that comes in its place.
• Scale the deployment; Kong then sends proxy requests to the new Pods and leave the short-circuited Pod out of the loop.
• Manually change the Pod health status in Kong using Kong’s Admin API.

These options highlight the fact that once a circuit is opened because of errors, there is no way for Kong to close the circuit again. Manual intervention is necessary so that a Pod can again handle requests. You can use active health-check, where each instance of Kong actively probes Pods to check if they are healthy.

Deploy an upstream HTTP application

Create a Kubernetes service setup a httpbin service in the cluster and proxy it.

kubectl apply -f https://docs.konghq.com/assets/kubernetes-ingress-controller/examples/httpbin-service.yaml

The results should look like this:

service/httpbin created
deployment.apps/httpbin created

Add routing configuration

1. Expose the service outside the Kubernetes cluster.

   Gateway API

   Ingress

   echo '
   apiVersion: gateway.networking.k8s.io/v1
   kind: HTTPRoute
   metadata:
     name: demo
     annotations:
       konghq.com/strip-path: "true"
   spec:
     parentRefs:
     - name: kong
     rules:
     - matches:
       - path:
           type: PathPrefix
           value: /test
       backendRefs:
       - name: httpbin
         kind: Service
         port: 80
   ' | kubectl apply -f -

   echo '
   apiVersion: networking.k8s.io/v1
   kind: Ingress
   metadata:
     name: demo
     annotations:
       konghq.com/strip-path: "true"
   spec:
     ingressClassName: kong
     rules:
     - http:
         paths:
         - path: /test
           pathType: Prefix
           backend:
             service:
               name: httpbin
               port:
                 number: 80
   ' | kubectl apply -f -

   The results should look like this:

   Gateway API

   Ingress

   httproute.gateway.networking.k8s.io/demo created

   ingress.networking.k8s.io/demo created

2. Test these endpoints.

    curl -i $PROXY_IP/test/status/200
   

   The results should look like this:

    HTTP/1.1 200 OK
    Content-Type: text/html; charset=utf-8
    Content-Length: 0
    Connection: keep-alive
    Server: gunicorn/19.9.0
    Access-Control-Allow-Origin: *
    Access-Control-Allow-Credentials: true
    X-Kong-Upstream-Latency: 1
    X-Kong-Proxy-Latency: 1
    Via: kong/3.4.2
   

   Observe the headers and you can see that Kong has proxied the request correctly.

Setup passive health checking

1. All health checks are done at Service-level and not Ingress-level. To configure Kong to short-circuit requests to a Pod if it throws 3 consecutive errors, add a KongUpstreamPolicy resource.

   echo '
   apiVersion: configuration.konghq.com/v1beta1
   kind: KongUpstreamPolicy
   metadata:
       name: demo-health-checking
   spec:
     healthchecks:
       passive:
         healthy:
           successes: 3
         unhealthy:
           httpFailures: 3
   ' | kubectl apply -f -
   

   The results should look like this:

   kongupstreampolicy.configuration.konghq.com/demo-health-checking created
   

2. Associate the KongUpstreamPolicy resource with httpbin service.

    kubectl patch svc httpbin -p '{"metadata":{"annotations":{"konghq.com/upstream-policy":"demo-health-checking"}}}'
   

   The results should look like this:

    service/httpbin patched
   

3. Test the Ingress rule by sending 2 requests that represent a failure from upstream and then send a request for 200. The requests with /status/500 simulate a failure from upstream. Send two requests with status/500.

    $ curl -i $PROXY_IP/test/status/500
    $ curl -i $PROXY_IP/test/status/500
   

   The results should look like this:

    HTTP/1.1 500 INTERNAL SERVER ERROR
    Content-Type: text/html; charset=utf-8
    Content-Length: 0
    Connection: keep-alive
    Server: gunicorn/19.9.0
    Access-Control-Allow-Origin: *
    Access-Control-Allow-Credentials: true
    X-Kong-Upstream-Latency: 1
    X-Kong-Proxy-Latency: 0
    Via: kong/3.4.2
   

   Send the third request with status/200

    $ curl -i $PROXY_IP/test/status/200
   

   The results should look like this:

    HTTP/1.1 200 OK
    Content-Type: text/html; charset=utf-8
    Content-Length: 0
    Connection: keep-alive
    Server: gunicorn/19.9.0
    Access-Control-Allow-Origin: *
    Access-Control-Allow-Credentials: true
    X-Kong-Upstream-Latency: 1
    X-Kong-Proxy-Latency: 0
    Via: kong/3.4.2
   

   Kong has not short-circuited because there were only two failures.

4. Send 3 requests to open the circuit, and then send a normal request. Send two requests with status/500.

    $ curl -i $PROXY_IP/test/status/500
    $ curl -i $PROXY_IP/test/status/500
    $ curl -i $PROXY_IP/test/status/500
   

   The results should look like this:

    HTTP/1.1 500 INTERNAL SERVER ERROR
    Content-Type: text/html; charset=utf-8
    Content-Length: 0
    Connection: keep-alive
    Server: gunicorn/19.9.0
    Access-Control-Allow-Origin: *
    Access-Control-Allow-Credentials: true
    X-Kong-Upstream-Latency: 1
    X-Kong-Proxy-Latency: 0
    Via: kong/3.4.2
   

   Send the fourth request with status/200

    curl -i $PROXY_IP/test/status/200
   

   The results should look like this:

    HTTP/1.1 503 Service Temporarily Unavailable
    Content-Type: application/json; charset=utf-8
    Connection: keep-alive
    Content-Length: 62
    X-Kong-Response-Latency: 0
    Server: kong/3.4.2
       
    {
      "message":"failure to get a peer from the ring-balancer"
    }%                  
   

   Kong returns a 503, indicating that the service is unavailable. Because there is only one Pod of httpbin service running in the cluster, and that is throwing errors, Kong does not proxy anymore requests. To get around this, you can use active health-check, where each instance of Kong actively probes Pods to check if they are healthy.

Setup active health checking

1. Update the KongUpstreamPolicy resource to use active health-checks.

   echo '
   apiVersion: configuration.konghq.com/v1beta1
   kind: KongUpstreamPolicy
   metadata:
       name: demo-health-checking
   spec:
     healthchecks:
       active:
         healthy:
           interval: 5
           successes: 3
         httpPath: /status/200
         type: http
         unhealthy:
           httpFailures: 1
           interval: 5
       passive:
         healthy:
           successes: 3
         unhealthy:
           httpFailures: 3
   ' | kubectl apply -f -
   

   The results should look like this:

   kongupstreampolicy.configuration.konghq.com/demo-health-checking configured
   

   This configures Kong to actively probe /status/200 every 5 seconds. If a Pod is unhealthy from Kong’s perspective, 3 successful probes changes the status of the Pod to healthy and Kong again starts to forward requests to that Pod. Wait 15 seconds for the pod to be marked as healthy before continuing.

2. Test the Ingress rule.

    curl -i $PROXY_IP/test/status/200
   

   The results should look like this:

    HTTP/1.1 200 OK
    Content-Type: text/html; charset=utf-8
    Content-Length: 0
    Connection: keep-alive
    Server: gunicorn/19.9.0
    Access-Control-Allow-Origin: *
    Access-Control-Allow-Credentials: true
    X-Kong-Upstream-Latency: 1
    X-Kong-Proxy-Latency: 1
    Via: kong/3.4.2
   

3. Trip the circuit again by sending three requests that returns the status 500 from httpbin.

   $ curl -i $PROXY_IP/test/status/500
   $ curl -i $PROXY_IP/test/status/500
   $ curl -i $PROXY_IP/test/status/500
   

   When you send the requests, it fails for about 15 seconds, the duration for active health checks to re-classify the httpbin Pod as healthy again.

   $ curl -i $PROXY_IP/test/status/200
   

   The results should look like this:

   HTTP/1.1 503 Service Temporarily Unavailable
   Content-Type: application/json; charset=utf-8
   Connection: keep-alive
   Content-Length: 62
   X-Kong-Response-Latency: 0
   Server: kong/3.4.2
       
   {
     "message":"failure to get a peer from the ring-balancer"
   }%                  
   

4. Send the requests after 15 seconds or so.

   curl -i $PROXY_IP/test/status/200
   

   The results should look like this:

   HTTP/1.1 200 OK
   Content-Type: text/html; charset=utf-8
   Content-Length: 0
   Connection: keep-alive
   Server: gunicorn/19.9.0
   Access-Control-Allow-Origin: *
   Access-Control-Allow-Credentials: true
   X-Kong-Upstream-Latency: 1
   X-Kong-Proxy-Latency: 1
   Via: kong/3.4.2
   

Read more about health-checks and circuit breaker in Kong’s documentation.