Deploying

PipelineBot is a Probot application that responds to events from GitHub and triggers tasks (workflows) based on actions in the application and check status events from GitHub.

It requires storage providers to be implemented as well as deployment infrastructure to deploy the code. The following deployment targets are implemented already to get you started quickly:

Deployment targets

• Firebase
• Kubernetes

Building A deployment target

Building a deployment target requires implementing a set of services defined in @deliverybot/core and then providing an HTTP server implementation that allows Deliverybot to receive HTTP traffic.

The @deliverybot/run package provides a local implementation for spinning up services and the HTTP interface for local development. Use this package as a starting point.

The load function exported from @deliverybot/core takes the services you implement along with the apps which provide the main functionality to give back a probot and express object to wire up to your infrastructure. A basic example of booting Deliverybot:

import { load, Options, localServices } from "@deliverybot/core";
import { apps } from "@deliverybot/app";

const options: Options = { ... };
load(localServices(), apps, options);

The main thing that you will need to replace when deploying the application is the services implementation. The services file provides the interfaces required by Deliverybot for your infrastructure.

Extending functionality

As mentioned the apps argument is where all of the functionality is wired up. You can easily extend Deliverybot with your own implementations and services by adding an app into the load array.

@deliverybot/core provides information and documentation on wiring up a new app for providing additional functionality.