Integrate Aha! with third-party services

This library provides integrations between Aha! and third-party services in response to user interface events within Aha! It is based on the Github-Services library.

This library supports integrations written in Ruby.

Each third-party service is represented by at least three files:

1. Implementation of the integration in lib/services/<service-name>.rb.
2. User documentation for configuring and using the integration in docs/<service-name>.md.
3. Unit test for the integration in spec/servces/<service-name>_spec.rb.

To add support for a new service, add these three files, implement and test your service, then submit a pull request.

Contact [email protected] for help using this code.

Service events

This service code is called in response to events triggered by the Aha! user interface. Each event is accompanied by a payload that describes the particular event in detail. The possible events are:

• installed - occurs when the service is installed by an Aha! user. This event can be triggered by the Test connection button in the Aha! integrations UI, but may also occur at other times. The event may occur many times for a service and must be idempotent. There is no payload.
• create_feature - occurs when a user chooses Send to in the Aha! UI and no integration fields exist for the feature. The payload contains complete information about the feature. See spec/fixtures/create_feature_event.json for an example.
• update_feature - occurs when a user chooses Update to in the Aha! UI and integration fields already exist for the feature. The payload contains complete information about the feature. See spec/fixtures/update_feature_event.json for an example.
• create_release - occurs before create_feature if there are no integration fields for the release the feature belongs to. The payload contains complete information about the release. See spec/fixtures/update_release_event.json for an example.
• update_release - occurs when there are changes to a release which has integration fields. The payload contains complete information about the release. See spec/fixtures/update_release_event.json for an example.
• webhook - occurs when a POST call to a webhook is received by Aha! The payload contains the body of the POST. This event can be used to make the integration bi-directional and react to changes happening in the integrated system.

Note that you can use the ngrok web interface to see the contents of the payloads being sent to your service for each of these events.

Configuration

Each service has two types of configuration information, each persistently stored and available via a Hashie::Mash. Each mash can store Ruby scalar values String, Symbol, Numeric, NilClass, TrueClass, FalseClass, Date, Time, DateTime as well as Array and Hash, but should not be used for other objects.

• meta_data - this mash is solely for the use of the integration to store any configuration information it needs to persist between events. Changes to the meta_data will only be saved after the installed event.
• data - this mash stores the user entered configuration information, corresponding to the fields in the Integrations user interface. It should not be modified by the integration code.

Integration fields

Each Aha! release, feature and requirement can have integration fields which track the relationship between that object and an integration. It is up to the integration implementation to decide how to use the integration fields, Aha! simply provides a key/value store that is per-record-per-integration.

When an object is created in a third party system, the integration code can use the Aha! API to store integration fields for the object. A typical pattern is to store the integration fields like:

api.create_integration_fields('features', feature.id, data.integration_id,
  {id: issue.id, key: issue[:key]})

That is, api.create_integration_fields(resource_type, resource_id, integration_id, data_hash)

The payload for each feature and release event will contain the integration fields that have previously been set.

Running an integration in development

You can run the integration in development by using a proxy server running on your development machine that passes messages between Aha! and the service code. Your development machine must be accessible to the Internet so that Aha! can send messages to it asynchronously. One way to do this is using ngrok.

The proxy works by sending HTTP messages from Aha! to your development machine to get configuration and remotely access the service. This allows you to rapidly develop and troubleshoot the service code using data from your live Aha! instance.

1. Run the proxy server on your local development machine:

   bundle
   PORT=4567 foreman start
   

2. In another terminal window start a tunnel to your running proxy server so that Aha! can access it, e.g. using ngrok:

   ngrok http 4567
   

3. Copy the HTTPS version of the Forwarding address, e.g. https://ba4a410.ngrok.com.

4. You can test this URL by loading https://ba4a410.ngrok.com/configuration in your browser.

5. Make the special "Development Proxy" integration visible in the Aha! UI by logging into your Aha! account and going to any existing integration. Add ?development=true to the URL and load the page. You will see a new integration appear named "Development Proxy".

6. Configure the development proxy settings. After you enter the proxy server URL go got in step 3 you will be able to choose from a list of the services that are running on your local machine.

7. Send features and releases to the "Development Proxy" and they will be send to your remotely running code.

The proxy server will auto reload when you save a file