User guide: Authenticated rate limiting (with Envoy Dynamic Metadata)
Provide Envoy with dynamic metadata about the external authorization process to be injected into the rate limiting filter.
Authorino features in this guide:
- Dynamic response → Response wrappers → Envoy Dynamic Metadata
- Dynamic response → JSON injection
- Identity verification & authentication → API key
Dynamic JSON objects built out of static values and values fetched from the [Authorization JSON](./../architecture.md#the-authorization-json) can be wrapped to be returned to the reverse-proxy as Envoy Well Known Dynamic Metadata content. Envoy can use those to inject data returned by the external authorization service into the other filters, such as the rate limiting filter.
Check out as well the user guides about [Injecting data in the request](./injecting-data.md) and [Authentication with API keys](./api-key-authentication.md).
For further details about Authorino features in general, check the [docs](./../features.md).
Requirements
- Kubernetes server
Create a containerized Kubernetes server locally using Kind:
1. Install the Authorino Operator
curl -sL https://raw.githubusercontent.com/Kuadrant/authorino-operator/main/utils/install.sh | bash -s
2. Deploy the Talker API
The Talker API is just an echo API, included in the Authorino examples. We will use it in this guide as the service to be protected with Authorino.
kubectl apply -f https://raw.githubusercontent.com/kuadrant/authorino-examples/main/talker-api/talker-api-deploy.yaml
3. Deploy Authorino
kubectl apply -f -<<EOF
apiVersion: operator.authorino.kuadrant.io/v1beta1
kind: Authorino
metadata:
name: authorino
spec:
listener:
tls:
enabled: false
oidcServer:
tls:
enabled: false
EOF
The command above will deploy Authorino as a separate service (as opposed to a sidecar of the protected API and other architectures), in namespaced
reconciliation mode, and with TLS termination disabled. For other variants and deployment options, check out the Getting Started section of the docs, the Architecture page, and the spec for the Authorino
CRD in the Authorino Operator repo.
4. Deploy Limitador
Limitador is a lightweight rate limiting service that can be used with Envoy.
On this bundle, we will deploy Limitador pre-configured to limit requests to the talker-api
domain up to 5 requests per interval of 60 seconds per user_id
. Envoy will be configured to recognize the presence of Limitador and activate it on requests to the Talker API.
kubectl apply -f https://raw.githubusercontent.com/kuadrant/authorino-examples/main/limitador/limitador-deploy.yaml
5. Setup Envoy
The following bundle from the Authorino examples (manifest referred in the command below) is to apply Envoy configuration and deploy Envoy proxy, that wire up the Talker API behind the reverse-proxy and external authorization with the Authorino instance.
For details and instructions to setup Envoy manually, see Protect a service > Setup Envoy in the Getting Started page. For a simpler and straightforward way to manage an API, without having to manually install or configure Envoy and Authorino, check out Kuadrant.
kubectl apply -f https://raw.githubusercontent.com/kuadrant/authorino-examples/main/envoy/envoy-notls-deploy.yaml
The bundle also creates an Ingress
with host name talker-api-authorino.127.0.0.1.nip.io
, but if you are using a local Kubernetes cluster created with Kind, you need to forward requests on port 8000 to inside the cluster in order to actually reach the Envoy service:
6. Create the AuthConfig
kubectl apply -f -<<EOF
apiVersion: authorino.kuadrant.io/v1beta2
kind: AuthConfig
metadata:
name: talker-api-protection
spec:
hosts:
- talker-api-authorino.127.0.0.1.nip.io
authentication:
"friends":
apiKey:
selector:
matchLabels:
group: friends
credentials:
authorizationHeader:
prefix: APIKEY
response:
success:
dynamicMetadata:
"rate-limit":
json:
properties:
"username":
selector: auth.identity.metadata.annotations.auth-data\/username
key: ext_auth_data # how this bit of dynamic metadata from the ext authz service is named in the Envoy config
EOF
An annotation auth-data/username
will be read from the Kubernetes Secret
s storing valid API keys and passed as dynamic metadata { "ext_auth_data": { "username": «annotations.auth-data/username» } }
.
Check out the docs for information about the common feature JSON paths for reading from the Authorization JSON.
7. Create a couple of API keys
For user John:
kubectl apply -f -<<EOF
apiVersion: v1
kind: Secret
metadata:
name: api-key-1
labels:
authorino.kuadrant.io/managed-by: authorino
group: friends
annotations:
auth-data/username: john
stringData:
api_key: ndyBzreUzF4zqDQsqSPMHkRhriEOtcRx
type: Opaque
EOF
For user Jane:
kubectl apply -f -<<EOF
apiVersion: v1
kind: Secret
metadata:
name: api-key-2
labels:
authorino.kuadrant.io/managed-by: authorino
group: friends
annotations:
auth-data/username: jane
stringData:
api_key: 7BNaTmYGItSzXiwQLNHu82+x52p1XHgY
type: Opaque
EOF
8. Consume the API
As John:
curl -H 'Authorization: APIKEY ndyBzreUzF4zqDQsqSPMHkRhriEOtcRx' http://talker-api-authorino.127.0.0.1.nip.io:8000/hello
# HTTP/1.1 200 OK
Repeat the request a few more times within the 60-second time window, until the response status is 429 Too Many Requests
.
While the API is still limited to John, send requests as Jane:
curl -H 'Authorization: APIKEY 7BNaTmYGItSzXiwQLNHu82+x52p1XHgY' http://talker-api-authorino.127.0.0.1.nip.io:8000/hello
# HTTP/1.1 200 OK
Cleanup
If you have started a Kubernetes cluster locally with Kind to try this user guide, delete it by running:
Otherwise, delete the resources created in each step:
kubectl delete secret/api-key-1
kubectl delete secret/api-key-2
kubectl delete authconfig/talker-api-protection
kubectl delete authorino/authorino
kubectl delete -f https://raw.githubusercontent.com/kuadrant/authorino-examples/main/envoy/envoy-notls-deploy.yaml
kubectl delete -f https://raw.githubusercontent.com/kuadrant/authorino-examples/main/talker-api/talker-api-deploy.yaml
kubectl delete -f https://raw.githubusercontent.com/kuadrant/authorino-examples/main/limitador/limitador-deploy.yaml
To uninstall the Authorino Operator and manifests (CRDs, RBAC, etc), run: