Menu Close

Chapter 10. Security

10.1. Configuring JSON Web Token authentication for Knative services

After the Service Mesh integration with OpenShift Serverless and Kourier has been configured on your cluster, you can enable JSON Web Token (JWT) authentication for your Knative services.

10.1.1. Enabling sidecar injection for a Knative service

You can add the sidecar.istio.io/inject="true" annotation to a Knative service to enable sidecar injection for that service.

Important

Adding sidecar injection to pods in system namespaces, such as knative-serving and knative-serving-ingress, is not supported when Kourier is enabled.

If you require sidecar injection for pods in these namespaces, see the OpenShift Serverless documentation on Integrating Service Mesh with OpenShift Serverless natively.

Prerequisites

  • You have installed the OpenShift Serverless Operator and Knative Serving.
  • Install the OpenShift CLI (oc).
  • You have created a project or have access to a project with the appropriate roles and permissions to create applications and other workloads in OpenShift Container Platform.

Procedure

  1. Add the sidecar.istio.io/inject="true" annotation to your Service resource:

    Example service

    apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: <service_name>
spec:
  template:
    metadata:
      annotations:
        sidecar.istio.io/inject: "true" 1
        sidecar.istio.io/rewriteAppHTTPProbers: "true" 2
...

    1
    Add the sidecar.istio.io/inject="true" annotation.
    2
    You must set the annotation sidecar.istio.io/rewriteAppHTTPProbers: "true" in your Knative service as OpenShift Serverless versions 1.14.0 and higher use an HTTP probe as the readiness probe for Knative services by default.

  2. Apply your Service resource YAML file: 

    $ oc apply -f <filename>

10.1.2. Using JSON Web Token authentication with Service Mesh 2.x and OpenShift Serverless

You can use the following procedure to enable using JSON Web Token authentication with Service Mesh 2.x and OpenShift Serverless.

Prerequisites

  • You have installed the OpenShift Serverless Operator and Knative Serving.
  • Install the OpenShift CLI (oc).
  • You have created a project or have access to a project with the appropriate roles and permissions to create applications and other workloads in OpenShift Container Platform.

Procedure

  1. Create a RequestAuthentication resource in each serverless application namespace that is a member in the ServiceMeshMemberRoll object: 

    apiVersion: security.istio.io/v1beta1
kind: RequestAuthentication
metadata:
  name: jwt-example
  namespace: <namespace>
spec:
  jwtRules:
  - issuer: testing@secure.istio.io
    jwksUri: https://raw.githubusercontent.com/istio/istio/release-1.8/security/tools/jwt/samples/jwks.json

  2. Apply the RequestAuthentication resource: 

    $ oc apply -f <filename>

  3. Allow access to the RequestAuthenticaton resource from system pods for each serverless application namespace that is a member in the ServiceMeshMemberRoll object, by creating the following AuthorizationPolicy resource: 

    apiVersion: security.istio.io/v1beta1
kind: AuthorizationPolicy
metadata:
  name: allowlist-by-paths
  namespace: <namespace>
spec:
  action: ALLOW
  rules:
  - to:
    - operation:
        paths:
        - /metrics 1
        - /healthz 2
    1
    The path on your application to collect metrics by system pod.
    2
    The path on your application to probe by system pod.

  4. Apply the AuthorizationPolicy resource: 

    $ oc apply -f <filename>

  5. For each serverless application namespace that is a member in the ServiceMeshMemberRoll object, create the following AuthorizationPolicy resource: 

    apiVersion: security.istio.io/v1beta1
kind: AuthorizationPolicy
metadata:
  name: require-jwt
  namespace: <namespace>
spec:
  action: ALLOW
  rules:
  - from:
    - source:
       requestPrincipals: ["testing@secure.istio.io/testing@secure.istio.io"]

  6. Apply the AuthorizationPolicy resource: 

    $ oc apply -f <filename>

Verification

  1. If you try to use a curl request to get the Knative service URL, it is denied:

    Example command

    $ curl http://hello-example-1-default.apps.mycluster.example.com/

    Example output

    RBAC: access denied

  2. Verify the request with a valid JWT.

    1. Get the valid JWT token: 

      $ TOKEN=$(curl https://raw.githubusercontent.com/istio/istio/release-1.8/security/tools/jwt/samples/demo.jwt -s) && echo "$TOKEN" | cut -d '.' -f2 - | base64 --decode -

    2. Access the service by using the valid token in the curl request header: 

      $ curl -H "Authorization: Bearer $TOKEN"  http://hello-example-1-default.apps.example.com

      The request is now allowed:

      Example output

      Hello OpenShift!

10.1.3. Using JSON Web Token authentication with Service Mesh 1.x and OpenShift Serverless

You can use the following procedure to enable using JSON Web Token authentication with Service Mesh 1.x and OpenShift Serverless.

Prerequisites

  • You have installed the OpenShift Serverless Operator and Knative Serving.
  • Install the OpenShift CLI (oc).
  • You have created a project or have access to a project with the appropriate roles and permissions to create applications and other workloads in OpenShift Container Platform.

Procedure

  1. Create a policy in a serverless application namespace which is a member in the ServiceMeshMemberRoll object, that only allows requests with valid JSON Web Tokens (JWT):

    Important

    The paths /metrics and /healthz must be included in excludedPaths because they are accessed from system pods in the knative-serving namespace. 

    apiVersion: authentication.istio.io/v1alpha1
kind: Policy
metadata:
  name: default
  namespace: <namespace>
spec:
  origins:
  - jwt:
      issuer: testing@secure.istio.io
      jwksUri: "https://raw.githubusercontent.com/istio/istio/release-1.6/security/tools/jwt/samples/jwks.json"
      triggerRules:
      - excludedPaths:
        - prefix: /metrics 1
        - prefix: /healthz 2
  principalBinding: USE_ORIGIN
    1
    The path on your application to collect metrics by system pod.
    2
    The path on your application to probe by system pod.

  2. Apply the Policy resource: 

    $ oc apply -f <filename>

Verification

  1. If you try to use a curl request to get the Knative service URL, it is denied: 

    $ curl http://hello-example-default.apps.mycluster.example.com/

    Example output

    Origin authentication failed.

  2. Verify the request with a valid JWT.

    1. Get the valid JWT token: 

      $ TOKEN=$(curl https://raw.githubusercontent.com/istio/istio/release-1.6/security/tools/jwt/samples/demo.jwt -s) && echo "$TOKEN" | cut -d '.' -f2 - | base64 --decode -

    2. Access the service by using the valid token in the curl request header: 

      $ curl http://hello-example-default.apps.mycluster.example.com/ -H "Authorization: Bearer $TOKEN"

      The request is now allowed:

      Example output

      Hello OpenShift!

10.2. Configuring a custom domain for a Knative service

Knative services are automatically assigned a default domain name based on your cluster configuration. For example, <service_name>.<namespace>.example.com.

You can customize the domain for your Knative service by mapping a custom domain name that you own to a Knative service, by creating a DomainMapping resource for the service. You can also create multiple DomainMapping resources to map multiple domains and subdomains to a single service.

10.2.1. Creating a custom domain mapping

To map a custom domain name to a custom resource (CR), you must create a DomainMapping CR that maps to an Addressable target CR, such as a Knative service or a Knative route.

Prerequisites

  • The OpenShift Serverless Operator and Knative Serving are installed on your cluster.
  • Install the OpenShift CLI (oc).
  • You have created a project or have access to a project with the appropriate roles and permissions to create applications and other workloads in OpenShift Container Platform.

  • You have created a Knative service and control a custom domain that you want to map to that service.

    Note

    Your custom domain must point to the IP address of the OpenShift Container Platform cluster.

Procedure

  1. Create a YAML file containing the DomainMapping CR in the same namespace as the target CR you want to map to: 

    apiVersion: serving.knative.dev/v1alpha1
kind: DomainMapping
metadata:
 name: <domain_name> 1
 namespace: <namespace> 2
spec:
 ref:
   name: <target_name> 3
   kind: <target_type> 4
   apiVersion: serving.knative.dev/v1
    1
    The custom domain name that you want to map to the target CR.
    2
    The namespace of both the DomainMapping CR and the target CR.
    3
    The name of the target CR to map to the custom domain.
    4
    The type of CR being mapped to the custom domain.

    Example service domain mapping

    apiVersion: serving.knative.dev/v1alpha1
kind: DomainMapping
metadata:
 name: example.com
 namespace: default
spec:
 ref:
   name: example-service
   kind: Service
   apiVersion: serving.knative.dev/v1

    Example route domain mapping

    apiVersion: serving.knative.dev/v1alpha1
kind: DomainMapping
metadata:
 name: example.com
 namespace: default
spec:
 ref:
   name: example-route
   kind: Route
   apiVersion: serving.knative.dev/v1

  2. Apply the DomainMapping CR as a YAML file: 

    $ oc apply -f <filename>

10.2.2. Creating a custom domain mapping by using the Knative CLI

You can use the kn CLI to create a DomainMapping custom resource (CR) that maps to an Addressable target CR, such as a Knative service or a Knative route.

The --ref flag specifies an Addressable target CR for domain mapping.

If a prefix is not provided when using the --ref flag, it is assumed that the target is a Knative service in the current namespace. The examples in the following procedure show the prefixes for mapping to a Knative service or a Knative route.

Prerequisites

  • The OpenShift Serverless Operator and Knative Serving are installed on your cluster.

  • You have created a Knative service or route, and control a custom domain that you want to map to that CR.

    Note

    Your custom domain must point to the DNS of the OpenShift Container Platform cluster.

  • You have installed the kn CLI tool.
  • You have created a project or have access to a project with the appropriate roles and permissions to create applications and other workloads in OpenShift Container Platform.

Procedure

  • Map a domain to a CR in the current namespace: 

    $ kn domain create <domain_mapping_name> --ref <target_name>

    Example command

    $ kn domain create example.com --ref example-service

  • Map a domain to a Knative service in a specified namespace: 

    $ kn domain create <domain_mapping_name> --ref <ksvc:service_name:service_namespace>

    Example command

    $ kn domain create example.com --ref ksvc:example-service:example-namespace

  • Map a domain to a Knative route: 

    $ kn domain create <domain_mapping_name> --ref <kroute:route_name>

    Example command

    $ kn domain create example.com --ref kroute:example-route

10.3. Using a custom TLS certificate for domain mapping

You can use an existing TLS certificate with a DomainMapping custom resource (CR) to secure the mapped service.

Prerequisites

10.3.1. Adding a custom TLS certificate to a DomainMapping CR

You can add an existing TLS certificate with a DomainMapping custom resource (CR) to secure the mapped service.

Prerequisites

  • You configured a custom domain for a Knative service and have a working DomainMapping CR.
  • You have a TLS certificate from your Certificate Authority provider or a self-signed certificate.
  • You have obtained the cert and key files from your Certificate Authority provider, or a self-signed certificate.
  • Install the OpenShift CLI (oc).

Procedure

  1. Create a Kubernetes TLS secret: 

    $ oc create secret tls <tls_secret_name> --cert=<path_to_certificate_file> --key=<path_to_key_file>

  2. Update the DomainMapping CR to use the TLS secret you have created: 

    apiVersion: serving.knative.dev/v1alpha1
kind: DomainMapping
metadata:
  name: <domain_name>
  namespace: <namespace>
spec:
  ref:
    name: <service_name>
    kind: Service
    apiVersion: serving.knative.dev/v1
# TLS block specifies the secret to be used
  tls:
    secretName: <tls_secret_name>

Verification

  1. Verify that the DomainMapping CR status is True, and that the URL column of the output shows the mapped domain with the scheme https: 

    $ oc get domainmapping <domain_name>

    Example output

    NAME                      URL                               READY   REASON
example.com               https://example.com               True

  2. Optional: If the service is exposed publicly, verify that it is available by running the following command: 

    $ curl https://<domain_name>

    If the certificate is self-signed, skip verification by adding the -k flag to the curl command.

10.4. Security configuration for Knative Kafka

In production, Kafka clusters are often secured using the TLS or SASL authentication methods. This section shows how to configure a Kafka channel to work against a protected Red Hat AMQ Streams cluster using TLS or SASL.

Note

If you choose to enable SASL, Red Hat recommends to also enable TLS.

10.4.1. Configuring TLS authentication

You can use the following procedure to configure TLS authentication for a Kafka channel.

Prerequisites

  • You have a Kafka cluster CA certificate stored as a .pem file.
  • You have a Kafka cluster client certificate and a key stored as .pem files.
  • Install the OpenShift CLI (oc).

Procedure

  1. Create the certificate files as secrets in your chosen namespace: 

    $ oc create secret -n <namespace> generic <kafka_auth_secret> \
  --from-file=ca.crt=caroot.pem \
  --from-file=user.crt=certificate.pem \
  --from-file=user.key=key.pem
    Important

    Use the key names ca.crt, user.crt, and user.key. Do not change them.

  2. Start editing the KnativeKafka custom resource: 

    $ oc edit knativekafka

  3. Reference your secret and the namespace of the secret: 

    apiVersion: operator.serverless.openshift.io/v1alpha1
kind: KnativeKafka
metadata:
  namespace: knative-eventing
  name: knative-kafka
spec:
  channel:
    authSecretName: <kafka_auth_secret>
    authSecretNamespace: <kafka_auth_secret_namespace>
    bootstrapServers: <bootstrap_servers>
    enabled: true
  source:
    enabled: true
    Note

    Make sure to specify the matching port in the bootstrap server.

    For example: 

    apiVersion: operator.serverless.openshift.io/v1alpha1
kind: KnativeKafka
metadata:
  namespace: knative-eventing
  name: knative-kafka
spec:
  channel:
    authSecretName: tls-user
    authSecretNamespace: kafka
    bootstrapServers: eventing-kafka-bootstrap.kafka.svc:9094
    enabled: true
  source:
    enabled: true

10.4.2. Configuring SASL authentication

You can use the following procedure to configure SASL authentication for a Kafka channel.

Prerequisites

  • You have a username and password for the Kafka cluster.
  • You have chosen the SASL mechanism to use, for example PLAIN, SCRAM-SHA-256, or SCRAM-SHA-512.
  • If TLS is enabled, you also need the ca.crt certificate file for the Kafka cluster.
  • Install the OpenShift CLI (oc).
Note

Red Hat recommends to enable TLS in addition to SASL.

Procedure

  1. Create the certificate files as secrets in your chosen namespace: 

    $ oc create secret --namespace <namespace> generic <kafka_auth_secret> \
  --from-file=ca.crt=caroot.pem \
  --from-literal=password="SecretPassword" \
  --from-literal=saslType="SCRAM-SHA-512" \
  --from-literal=user="my-sasl-user"
    Important

    Use the key names ca.crt, password, and saslType. Do not change them.

  2. Start editing the KnativeKafka custom resource: 

    $ oc edit knativekafka

  3. Reference your secret and the namespace of the secret: 

    apiVersion: operator.serverless.openshift.io/v1alpha1
kind: KnativeKafka
metadata:
  namespace: knative-eventing
  name: knative-kafka
spec:
  channel:
    authSecretName: <kafka_auth_secret>
    authSecretNamespace: <kafka_auth_secret_namespace>
    bootstrapServers: <bootstrap_servers>
    enabled: true
  source:
    enabled: true
    Note

    Make sure to specify the matching port in the bootstrap server.

    For example: 

    apiVersion: operator.serverless.openshift.io/v1alpha1
kind: KnativeKafka
metadata:
  namespace: knative-eventing
  name: knative-kafka
spec:
  channel:
    authSecretName: scram-user
    authSecretNamespace: kafka
    bootstrapServers: eventing-kafka-bootstrap.kafka.svc:9093
    enabled: true
  source:
    enabled: true

10.4.3. Configuring SASL authentication using public CA certificates

If you want to use SASL with public CA certificates, you must use the tls.enabled=true flag, rather than the ca.crt argument, when creating the secret. For example: 

$ oc create secret --namespace <namespace> generic <kafka_auth_secret> \
  --from-literal=tls.enabled=true \
  --from-literal=password="SecretPassword" \
  --from-literal=saslType="SCRAM-SHA-512" \
  --from-literal=user="my-sasl-user"

10.4.4. Additional resources