examples/knative/README.adoc

Knative Example (Apache Camel K)
================================

This example shows how Camel K can be used to connect Knative building blocks to create awesome applications.

A video version of this https://youtu.be/btf_e2GniXM[demo is available on YouTube].

It's assumed that both Camel K and Knative are properly installed (including Knative Build, Serving and Eventing) into the cluster.
Refer to the specific documentation to install and configure all components.

We're going to create two channels:
- messages
- words

The first channel will contain phrases, while the second one will contains the single words contained in the phrases.

To create the channels (they use the in-memory channel provisioner):

```
kubectl create -f messages-channel.yaml
kubectl create -f words-channel.yaml
```

We can now proceed to install all camel K integrations.

== Install a "Printer"

We'll install a Camel K integration that will print all words from the `words` channel.

Writing a "function" that does this is as simple as writing:

```
from('knative:channel/words')
  .convertBodyTo(String.class)
  .to('log:info')
```

You can run this integration by running:

```
kamel run printer.groovy
```

Under the hood, the Camel K operator does this:
- Understands that the integration is passive, meaning that it can be activated only using an external HTTP call (the knative consumer endpoint)
- Materializes the integration as a Knative autoscaling service, integrated in the Istio service mesh
- Adds a Knative Eventing `Subscription` that points to the autoscaling service

The resulting integration will be scaled to 0 when not used (if you wait ~5 minutes, you'll see it).

== Install a "Splitter"

We're now going to deploy a splitter, using the Camel core Split EIP. The splitter will take all messages from the `messages` channel,
split them and push the single words into the `words` channel.

The integration code is super simple:

```
from('knative:channel/messages')
  .split().tokenize(" ")
  .log('sending ${body} to words channel')
  .to('knative:channel/words')
```

Let's run it with:

```
kamel run splitter.groovy
```

This integration will be also materialized as a Knative autoscaling service, because the only entrypoint is passive (waits for a push notification).

== Install a "Feed"

We're going to feed this chain of functions using a timed feed like this:

```
from('timer:clock?period=3000')
  .setBody().constant("Hello World from Camel K")
  .to('knative:channel/messages')
  .log('sent message to messages channel')
```

Every 3 seconds, the integration sends a message to the Knative `messages` channel.

Let's run it with:

```
kamel run feed.groovy
```

This cannot be materialized into an autoscaling service, but the operator understands it automatically and maps it to a plain Kubernetes Deployment
(Istio sidecar will be injected).

== Playing around

If you've installed all the services, you'll find that the printer pod will print single words as they arrive from the feed (every 3 seconds, passing by the splitter function).

If you now stop the feed integration (`kamel delete feed`) you will notice that the other services (splitter and printer) will scale down to 0 in few minutes.

And if you reinstall the feed again (`kamel run feed.groovy`), the other integration will scale up again as soon as they receive messages (splitter first, then printer).

== Playing harder

You can also play with different kind of feeds. E.g. the following simple feed can be used to bind messages from Telegram to the system:

```
from('telegram:bots/<put-here-your-botfather-authorization>')
  .convertBodyTo(String.class)
  .to('log:info')
  .to('knative:channel/messages')
```

Now just send messages to your bot with the Telegram client app to see all single words appearing in the printer service.

You can also replace the printer with a Slack-based printer like:

```
from('knative:channel/words')
  .log('Received: ${body}')
  .to('slack:#camel-k-tests')


camel {
  components {
    slack {
      webhookUrl '<put-here-your-slack-incoming-webhook-url>'
    }
  }
}
```

Now the single words will be printed in the log but also forwarded to the
slack channel named `#camel-k-tests`.

You have infinite possibilities with Camel!