Control Headers and Routing

 4 minute read

This task demonstrates how to use a policy adapter to manipulate request headers and routing.

Before you begin

  • Set up Istio on Kubernetes by following the instructions in the Installation guide.

  • Follow the set-up instructions in the ingress task to configure an ingress using a gateway.

  • Customize the virtual service configuration for the httpbin service containing two route rules that allow traffic for paths /headers and /status:

    $ kubectl apply -f - <<EOF
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
  name: httpbin
spec:
  hosts:
  - "*"
  gateways:
  - httpbin-gateway
  http:
  - match:
    - uri:
        prefix: /headers
    - uri:
        prefix: /status
    route:
    - destination:
        port:
          number: 8000
        host: httpbin
EOF

Output-producing adapters

In this task, we are using a sample policy adapter keyval. In addition to a policy check result, this adapter returns an output with a single field called value. The adapter is configured with a lookup table, which it uses to populate the output value, or return NOT_FOUND error status if the input instance key is not present in the lookup table.

  1. Deploy the demo adapter:

    $ kubectl run keyval --image=gcr.io/istio-testing/keyval:release-1.1 --namespace istio-system --port 9070 --expose

  2. Enable the keyval adapter by deploying its template and configuration descriptors:

    ZipZip
    $ kubectl apply -f @samples/httpbin/policy/keyval-template.yaml@
$ kubectl apply -f @samples/httpbin/policy/keyval.yaml@

  3. Create a handler for the demo adapter with a fixed lookup table:

    $ kubectl apply -f - <<EOF
apiVersion: config.istio.io/v1alpha2
kind: handler
metadata:
  name: keyval
  namespace: istio-system
spec:
  adapter: keyval
  connection:
    address: keyval:9070
  params:
    table:
      jason: admin
EOF

  4. Create an instance for the handler with the user request header as a lookup key:

    $ kubectl apply -f - <<EOF
apiVersion: config.istio.io/v1alpha2
kind: instance
metadata:
  name: keyval
  namespace: istio-system
spec:
  template: keyval
  params:
    key: request.headers["user"] | ""
EOF

Request header operations

  1. Ensure the httpbin service is accessible through the ingress gateway:

    $ curl http://$INGRESS_HOST:$INGRESS_PORT/headers
{
  "headers": {
    "Accept": "*/*",
    "Content-Length": "0",
    ...
    "X-Envoy-Internal": "true"
  }
}

    The output should be the request headers as they are received by the httpbin service.

  2. Create a rule for the demo adapter:

    $ kubectl apply -f - <<EOF
apiVersion: config.istio.io/v1alpha2
kind: rule
metadata:
  name: keyval
  namespace: istio-system
spec:
  actions:
  - handler: keyval.istio-system
    instances: [ keyval ]
    name: x
  requestHeaderOperations:
  - name: user-group
    values: [ x.output.value ]
EOF

  3. Issue a new request to the ingress gateway with the header key set to value jason:

    $ curl -Huser:jason http://$INGRESS_HOST:$INGRESS_PORT/headers
{
  "headers": {
    "Accept": "*/*",
    "Content-Length": "0",
    "User": "jason",
    "User-Agent": "curl/7.58.0",
    "User-Group": "admin",
    ...
    "X-Envoy-Internal": "true"
  }
}

    Note the presence of the user-group header with the value derived from the rule application of the adapter. The expression x.output.value in the rule evaluates to the populated value field returned by the keyval adapter.

  4. Modify the rule to rewrite the URI path to a different virtual service route if the check succeeds:

    $ kubectl apply -f - <<EOF
apiVersion: config.istio.io/v1alpha2
kind: rule
metadata:
  name: keyval
  namespace: istio-system
spec:
  match: source.labels["istio"] == "ingressgateway"
  actions:
  - handler: keyval.istio-system
    instances: [ keyval ]
  requestHeaderOperations:
  - name: :path
    values: [ '"/status/418"' ]
EOF

  5. Repeat the request to the ingress gateway:

    $ curl -Huser:jason -I http://$INGRESS_HOST:$INGRESS_PORT/headers
HTTP/1.1 418 Unknown
server: istio-envoy
...

    Note that the ingress gateway changed the route after the rule application of the policy adapter. The modified request may use a different route and destination and is subject to the traffic management configuration.

    The modified request is not checked again by the policy engine within the same proxy. Therefore, we recommend to use this feature in gateways, so that the server-side policy checks take effect.

Cleanup

Delete the policy resources for the demo adapter:

$ kubectl delete rule/keyval handler/keyval instance/keyval adapter/keyval template/keyval -n istio-system
$ kubectl delete service keyval -n istio-system
$ kubectl delete deployment keyval -n istio-system

Complete the clean-up instructions in ingress task.