Skip to main content

Running Pipes

A Pipes can be run in two ways:

  1. On an Agent
  2. Through the CLI

Pipes are typically deployed on Agents. The latter method is useful for testing and troubleshooting.

Deploying a Pipe to an Agent​

This requires a running Server with at least one Agent.

Via the UI​

  1. Log in to the Web UI
  2. Go to Manage -> Pipes -> New Pipe
  3. Give the Pipe a name
  4. Edit (or paste) the Pipe definition, and save it
  5. Close the Pipe Editor
  6. On the Pipe view, click the Add button next to the available Agent(s)

The Server will prepare the Pipe payload for Agent, which will retrieve the Pipe Payload and start the Pipe. This usually happens within a few seconds.

Via the CLI​

On a terminal where the Server is running:

  1. Authenticate to the Server with hotrod login USERNAME (typically admin)
  2. Save the Pipe definition in a YAML file somewhere on disk
  3. Add the Pipe to the Server with hotrod pipes add -f /path/to/pipe.definition.yaml
  4. Deploy the Pipe to an Agent with hotrod agents update --add-pipe pipeName agentName

Through the CLI​

On a terminal with the Hotrod binary installed:

  1. Save the Pipe definition in a YAML file somewhere on disk
  2. Run the Pipe with hotrod pipes run -f /path/to/pipe.definition.yml

This method of executing a Pipe has some limitations. It runs the Pipe with an ad-hoc Pipe Runtime, which involves neither the Server nor an Agent. Any context variables used in the Pipe must be defined in the Pipe itself, or provided explicitly with the -c option. No Server or Agent context is available in this mode of execution.