eventing

Knative Events

Knative Events is a system which is designed to address a common need for cloud native development:

1. Services are loosely coupled during development and deployed independently on a variety of platforms (Kubernetes, VMs, SaaS or FaaS).
2. A producer can generate events before a consumer is listening, and a consumer can express an interest in an event or class of events that is not yet being produced.
3. Services can be connected to create new applications
   • without modifying producer or consumer.
   • with the ability to select a specific subset of events from a particular producer

The above concerns are consistent with the design goals of CloudEvents, a common specification for cross-service interoperability being developed by the CNCF Serverless WG.

Note: The Eventing repo is very much a work-in-progress. See the list of Known Issues with the current state, but feel free to report additional issues. Expect to see changes in all of the following areas in the next months:

1. The API model and terminology is still evolving.
2. The controller model for Buses and Sources are still being refined.
3. Test coverage is not up to snuff.
4. The release packaging is still being determined.

Once these concerns are resolved, we expect to work on creating a wider array of EventSources.

Installation

You can install Knative Eventing with the following command:

kubectl apply --filename https://storage.googleapis.com/knative-releases/eventing/latest/release.yaml

In addition to the core definitions, you'll need to install at least one EventSource and one Bus to have a useful cluster. See below for more information on how to install different versions of these components.

A note on naming: We track the CloudEvents nomenclature where possible. Much of the naming in this area is in flux, however and many of the needed terms are outside of the current scope of the specification.

Architecture

In order to subdivide the problem, we have split the server-side components into three abstractions:

Buses

Buses provide a k8s-native abstraction over message buses like NATS or Kafka. At this level, the abstraction is basically publish-subscribe; events are published to a Channel, and Subscriptions route that Channel to interested parties.

• Channel is a network endpoint which receives (and optionally persists) events using a Bus-specific implementation.
• Subscription connects events received on a Channel to an interested target, represented as a DNS name. There may be multiple Subscriptions on a single channel.
• Bus defines the adaptation layers needed to implement channels and subscriptions using a specific persistence strategy (such as delivery of events to a Kafka topic).

We currently have 3 buses implemented:

• Stub provides a zero-dependency in-memory transport.

  kubectl apply --filename https://storage.googleapis.com/knative-releases/eventing/latest/release-bus-stub.yaml
  kubectl apply --filename https://storage.googleapis.com/knative-releases/eventing/latest/release-clusterbus-stub.yaml

• Kafka uses an existing (user-provided) Kafka cluster for persistence.

  kubectl apply --filename https://storage.googleapis.com/knative-releases/eventing/latest/release-bus-kafka.yaml
  kubectl apply --filename https://storage.googleapis.com/knative-releases/eventing/latest/release-clusterbus-kafka.yaml

• GCP PubSub uses Google Cloud PubSub for message persistence.

  kubectl apply --filename https://storage.googleapis.com/knative-releases/eventing/latest/release-bus-gcppubsub.yaml
  kubectl apply --filename https://storage.googleapis.com/knative-releases/eventing/latest/release-clusterbus-gcppubsub.yaml

Sources

Sources provide a similar abstraction layer for provisioning data sources from outside Kubernetes and routing them to the cluster, expressed as a Feed (typically, to a Channel on a Bus, but could also be directly to another endpoint). Right now, we only have a few generic Sources, but we plan to add more interesting and specific Sources over time.

• Feed is a primitive object defining a connection between an EventType and the action (as an CloudEvents compatible HTTP endpoint).
• EventType and ClusterEventType describe a specific set of events with a common schema which are emitted by an EventSource. EventType is namespace-scoped, while ClusterEventTypes are installed by an administrator and available in all namespaces in your cluster.
• EventSource and ClusterEventSource describe an external system which may produce one or more EventTypes.

We currently have 3 sources implemented:

• K8sevents collects Kubernetes Events and presents them as CloudEvents.

  kubectl apply --filename https://storage.googleapis.com/knative-releases/eventing/latest/release-source-k8sevents.yaml

• GitHub collects pull request notifications and presents them as CloudEvents.

  kubectl apply --filename https://storage.googleapis.com/knative-releases/eventing/latest/release-source-github.yaml

• GCP PubSub collects events published to a GCP PubSub topic and presents them as CloudEvents.

  kubectl apply --filename https://storage.googleapis.com/knative-releases/eventing/latest/release-source-gcppubsub.yaml

Flows

Lastly, we have a higher-level abstraction called a Flow which bundles up the specification from the Source to the endpoint, optionally allowing you to choose the Channel and Bus which the event is routed over. (Otherwise, there is a default Bus used to provision a Channel.)

• Flow is the top-level user-facing concept in Eventing; it describes the desired path from an external Source of events to a destination that will react to the events. There is only a single type of Flow, which is installed by the core Knative Eventing install.

Supporting Libraries

In addition to the above, there is a golang library at pkg/event for handling CloudEvents and easily creating a container which processes CloudEvents delivered over HTTP. You can look at several of the samples for usage examples.

Getting Started

• Setup Knative Serving
• Install Eventing components
• Run samples

---

Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License.