Skip to content

Quickstart

This quickstart will take you through the fastest and easiest way to start using events with TriggerMesh's command line interface, tmctl. tmctl makes it easy to create, configure and run TriggerMesh on any machine that has Docker.

By the end of this quickstart you will have done the following:

  1. captured events an external service by polling it over HTTP. It generates text that will be used as the event payload.
  2. Transformed the event by adding a new attribute to it
  3. Routed the event to an external Target: the TriggerMesh Console
  4. Throughout the quickstart, you'll monitor everything happening inside the Broker with the watchcommand

Docker required

Please make sure you are able to run docker containers on your machine before proceeding.

Installation

Install tmctl

tmctl is the TriggerMesh command line interface (CLI) that makes it easy to work with events from the safety of your office (or couch).

TriggerMesh CLI can be installed from different sources: brew repository, pre-built binary, or compiled from source.

brew install triggermesh/cli/tmctl

Use the following one-liner to automatically download and install the CLI:

curl -sSfL https://raw.githubusercontent.com/triggermesh/tmctl/HEAD/hack/install.sh | sh

Alternatively, visit the releases page to manually download the latest version for your platform and make sure the path of the downloaded binary is configured in the PATH environment variable.

Visit the releases page to manually download the latest version for your platform and make sure the path of the downloaded binary is configured in the PATH environment variable.

The most recent version of the go compiler is recommended to build tmctl binary from source:

git clone git@github.com:triggermesh/tmctl.git
cd tmctl
go install

Make sure that the binary's location is configured in the PATH environment variable, like so:

export PATH="$HOME/go/bin:$PATH"

Verify that the binary has been successfully installed:

tmctl --help

To configure autocompletion for tmctl, please refer to the dedicated guide.

Create a Broker and send it an event

Create a Broker

Your Broker will reliably route and deliver events from producers to consumers:

tmctl create broker foo

Watch events sent to the Broker

Open a second terminal to display all events passing through the broker. Keep this second terminal open throughout the quickstart.

Type:

tmctl watch

Wait for this command to say complete and print Watching....

Send it an event

Back to the 1st terminal, and send an event to the Broker with JSON data:

tmctl send-event '{"hello":"world"}'

In the Watching...terminal, you should see the event received by your Broker, something like this:

tmctl % tmctl watch
2022/11/08 16:35:33 Watching...
☁️  cloudevents.Event
Context Attributes,
  specversion: 1.0
  type: triggermesh-local-event
  source: triggermesh-cli
  id: 93dbed08-fb66-43d7-811b-2725f0e1d74e
  time: 2022-11-08T15:38:35.507676Z
  datacontenttype: application/json
Data,
  {
    "hello": "world"
  }

Doesn't work?

Copy any output and PLEASE TELL US NOW .

Add an external event source

Poll an external HTTP endpoint for events

Let's create external source:

tmctl create source httppoller \
    --endpoint https://corporatebs-generator.sameerkumar.website/ \
    --eventType buzzword.phrase \
    --interval 20s \
    --method GET

With the eventType parameter we're saying that events from this source should be given the type buzzword.phrase. This can be used later on for routing.

Watch the events

In the second terminal that is Watching...the Broker, you should be regularly receiving some generated text:

2022/11/09 16:17:38 Watching...
☁️  cloudevents.Event
Context Attributes,
  specversion: 1.0
  type: buzzword.phrase
  source: local.foo-httppollersource
  id: c3cf14c5-8228-4366-8a33-c42c8e5847cb
  time: 2022-11-22T13:07:51.659171708Z
  datacontenttype: application/json
Data,
  {
    "phrase": "Rapidiously Brand Covalent Architectures"
  }

Route events from the Broker to the TriggerMesh Console

Create a target

Now let's send those events somewhere.

We're going to run a local service called TriggerMesh Console that is used to view events in the browser. Create a Target that points to this service and runs the necessary container with the following command:

tmctl create target \
    --name console \
    --from-image gcr.io/triggermesh/triggermesh-console:v0.0.1 \
    --source foo-httppollersource

After running this command, it should output a URL that you can open in your browser:

Listening on:     http://localhost:<port>

View events in TriggerMesh Console

Open this in your browser, and watch as a new event appears every 20 seconds:

If you still have tmctl watch running in a second terminal, you'll also see events there when they hit the Broker, before passing on to the Console.

Transform the event

Lets transform the incoming events before they are passed to the Console.

Create a transformation

Using TriggerMesh's Bumbleebee transformation component, we can easily modify an event as it passes through the TriggerMesh.

tmctl create transformation --target console <<EOF
data:
- operation: add
  paths:
  - key: new-field
    value: hello from Transformation!
EOF

This simple transformation adds a new key to the event's JSON payload. We're using the --target parameter to indicate that the transformed events should be passed along to the Console.

Send an event

Although you can wait till the HTTPPoller fetches another event, you can also simulate you own event with the following command:

tmctl send-event '{"hello":"triggermesh"}' --eventType buzzword.phrase

Notice how events displayed in the Console now include the additional field that was added by the transformation.

View your TriggerMesh configuration

Run the following command:

tmctl describe

As you can see, tmctl describe displays useful info about your current configuration. It lists all the sources, targets, triggers, and other components you've defined, and their properties.

~ % tmctl describe
Broker     Status
foo        online(http://localhost:53053)

Trigger                  Target                 Filter
foo-trigger-a73dcb65     foo-transformation     type is buzzword.phrase
foo-trigger-0a1c285d     console                type is foo-transformation.output

Transformation         EventTypes                    Status
foo-transformation     foo-transformation.output     online(http://localhost:53205)

Source                   Kind                 EventTypes          Status
foo-httppollersource     httppollersource     buzzword.phrase     online(http://localhost:53069)

Target      Kind                                                        Expected Events     Status
console     service (gcr.io/triggermesh/triggermesh-console:v0.0.1)     *                   online(http://localhost:53098)

Next steps

Now that you know what a source, target, and transformation are, you can start doing more things! Like creating your own Trigger, or adding new sources and targets that make sense to you.

create your first Trigger:

AWS users:

Azure users:

Google Cloud users:

Kafka users:

HTTP users:

Kubernetes users:

  • try exporting your local configuration as a Kubernetes manifest by using tmctl dump, and deploying to a cluster. You can learn more about it in the dedicated guide.

Clean up

After running through this quickstart guide, you'll have a number of containers running on Docker.

You can stop all TriggerMesh-related containers with:

tmctl stop

To delete you local configuration (all sources, targets, brokers, transformations) and stop all containers, you can try:

tmctl delete --broker foo

Replace foo with the name of your broker. Be careful, you can't go back once you've deleted everything.