About sinks

When you create an event source, you can specify a sink where events are sent to from the source. A sink is an Addressable or a Callable resource that can receive incoming events from other resources. Knative Services, Channels, and Brokers are all examples of sinks.

Addressable objects receive and acknowledge an event delivered over HTTP to an address defined in their status.address.url field. As a special case, the core Kubernetes Service object also fulfils the Addressable interface.

Callable objects are able to receive an event delivered over HTTP and transform the event, returning 0 or 1 new events in the HTTP response. These returned events may be further processed in the same way that events from an external event source are processed.

Sink as a parameter

Sink is used as a reference to an object that resolves to a URI to use as the sink.

A sink definition supports the following fields:

Field	Description	Required or optional
ref	This points to an Addressable.	Required if not using uri
ref.apiVersion	API version of the referent.	Required if using ref
ref.kind	Kind of the referent.	Required if using ref
ref.namespace	Namespace of the referent. If omitted this defaults to the object holding it.	Optional
ref.name	Name of the referent.	Required if using ref
uri	This can be an absolute URL with a non-empty scheme and non-empty host that points to the target or a relative URI. Relative URIs are resolved using the base URI retrieved from Ref.	Required if not using ref

Note

At least one of ref or uri is required. If both are specified, uri is resolved into the URL from the Addressable ref result.

Sink parameter example

Given the following YAML, if ref resolves into "http://mysink.default.svc.cluster.local", then uri is added to this resulting in "http://mysink.default.svc.cluster.local/extra/path".

apiVersion: sources.knative.dev/v1
kind: SinkBinding
metadata:
  name: bind-heartbeat
spec:
  ...
  sink:
    ref:
      apiVersion: v1
      kind: Service
      namespace: default
      name: mysink
    uri: /extra/path

Contract

This results in the K_SINK environment variable being set on the subject as "http://mysink.default.svc.cluster.local/extra/path".

Using custom resources as sinks

To use a Kubernetes custom resource (CR) as a sink for events, you must:

1. Make the CR Addressable. You must ensure that the CR contains a status.address.url. For more information, see the spec for Addressable resources.

2. Create an Addressable-resolver ClusterRole to obtain the necessary RBAC rules for the sink to receive events.

   For example, you can create a kafkasinks-addressable-resolver ClusterRole to allow get, list, and watch access to KafkaSink objects and statuses:

   kind: ClusterRole
   apiVersion: rbac.authorization.k8s.io/v1
   metadata:
     name: kafkasinks-addressable-resolver
     labels:
       kafka.eventing.knative.dev/release: devel
       duck.knative.dev/addressable: "true"
   # Do not use this role directly. These rules will be added to the "addressable-resolver" role.
   rules:
     - apiGroups:
         - eventing.knative.dev
       resources:
         - kafkasinks
         - kafkasinks/status
       verbs:
         - get
         - list
         - watch
   

Filtering events sent to sinks by using Triggers

You can connect a Trigger to a sink, so that events are filtered before they are sent to the sink. A sink that is connected to a Trigger is configured as a subscriber in the Trigger resource spec.

For example:

apiVersion: eventing.knative.dev/v1
kind: Trigger
metadata:
  name: <trigger-name>
spec:
...
  subscriber:
    ref:
      apiVersion: eventing.knative.dev/v1alpha1
      kind: KafkaSink
      name: <kafka-sink-name>

Where;

• <trigger-name> is the name of the Trigger being connected to the sink.
• <kafka-sink-name> is the name of a KafkaSink object.

Specifying sinks using the kn CLI --sink flag

When you create an event-producing CR by using the Knative (kn) CLI, you can specify a sink where events are sent to from that resource, by using the --sink flag.

The following example creates a SinkBinding that uses a Service, http://event-display.svc.cluster.local, as the sink:

kn source binding create bind-heartbeat \
  --namespace sinkbinding-example \
  --subject "Job:batch/v1:app=heartbeat-cron" \
  --sink http://event-display.svc.cluster.local \
  --ce-override "sink=bound"

The svc in http://event-display.svc.cluster.local determines that the sink is a Knative Service. Other default sink prefixes include Channel and Broker.

Tip

You can configure which resources can be used with the --sink flag for kn CLI commands by customizing kn.

Supported third-party sink types

Name	Maintainer	Description
JobSink	Knative	Trigger long-running background jobs
KafkaSink	Knative	Send events to a Kafka topic
RedisSink	Knative	Send events to a Redis Stream