This repository provides a foundation for supporting webhook automation with Atlassian's Jira Service Desk. The aim of the framework is to provide a high quality foundation for providing webhook extensions to request types on Service Desk while reducing the amount of code needed to be written. The framework provides the core webhook handling code plus common functions for interacting with Service Desk and LDAP, thus allowing the handlers for each request type to focus on the "business logic" required.
The framework can be triggered by the following actions:
- Creation of a ticket (
CREATE) - Comment added to a ticket (
COMMENT) - Organization change on a ticket (
ORGCHANGE) - Status change on a ticket (
TRANSITION) - Assignment change on a ticket (
ASSIGNMENT) - Generic Jira webhook (
JIRAHOOK)
Each request type has its own code file. The main code loads the code file appropriate to the ticket being handled.
Since the webhooks can be generic across all of the issue/request types, each code file has a CAPABILITIES section that the main code queries in order to determine whether or not the code can support the desired action, e.g.:
CAPABILITIES = [ "CREATE", "ASSIGNMENT" ]
The sample test handler rt_example_handler.py lists all of the supported transition names. Handlers used by Linaro can be found in https://github.com/linaro-its/sd-webhook-handlers.
There are a few options for running the framework in production:
Both Chalice and Zappa are serverless options, but Chalice requires code changes in order to work.
To make it easier to use the framework with WSGI, this repository includes the files and configuration required to build a Docker container for running everything. See WSGI for more details about how to use the Docker container, including how to debug the code if required.
To use the framework with Zappa, there are some additional steps required. To simplify the steps required, a sample repo has been created that shows how to use this framework repo with the sample handlers repo. See Linaro SD Webhook for more details.
There is a commented configuration file - configuration.sample.jsonc - which needs to be copied as configuration.jsonc and then edited. This configuration file controls how the framework behaves to meet your specific needs, e.g. Service Desk bot account authentication, which handler file to use for which request type, etc. It is safe to leave the comments in the file - the framework copes with them being there when the file is read in.
If you are self-hosting Service Desk and you have Automation for Jira installed, use the instructions for Service Desk Cloud configuration, otherwise follow the instructions for "Service Desk (Server) webhooks" and "Jira (server) webhook".
In the text below, <base URL> is the URL that is mapped onto the framework web service. If the framework is being run on the same server as Jira Service Desk, this could be (for example) http://localhost:8000.
The Service Desk (Server) webhooks are created through Project Settings > Automation for a given SD project. Generally, the following rules should be configured, depending on the needs of the handlers:
-
When: Issue created
- Then: Webhook
- URL:
<base URL>/create - Include payload in request body
-
When: Comment added
- Then: Webhook
- URL:
<base URL>/comment - Include payload in request body
-
When: Organizations added to issue
- Then: Webhook
- URL:
<base URL>/org-change - Include payload in request body
You can use the IF clause to only fire when request types matching those supported by the webhook are affected. This is done by adding an Issue matches clause with JQL like this:
"Customer Request Type" = "Name of the request type"
or
Issue matches status = "Open"
Note that if you change the name of the request type, you must update any JQL clauses that reference it to match otherwise it will not fire. The clause does not get updated automatically.
Note that the framework will work out which request type has been used, so it is safe to omit the IF clause, with the understanding that the framework will be called for all issues created and commented on.
Create a WebHook in Jira with the following settings:
- URL:
<base URL>/jira-hook - Events: Issue>updated
The framework looks at the data sent by Jira and determines if it was an assignment or transition that triggered the event and acts accordingly. Note that a Jira webhook is used for transitions rather than the Service Desk "Status changed" event because it is then possible to track the before and after states.
If you want to be more flexible with the Jira webhook, e.g. fire on reasons other than assignment or transition, use the JIRAHOOK capability. The framework will trigger the handler code for a generic hook if the reason is not assignment or transition.
For all webhook configurations, you can specify a JQL query to restrict the webhook to appropriate projects and request types in order to ensure that the webhook only fires when appropriate. To filter on request types, use Customer Request Type.
If you have "Automation for Jira" installed on Service Desk (Server), you can configure these rules via Project settings > Project automation. For Service Desk (Cloud), you can configure these rules via Project settings > Automation.
Create automation rules for each of the following triggers:
- Issue created
- Issue commented
- Field value changed
- Fields to monitor for changes: Organizations
- For: All issue operations
- Issue transitioned
- Issue assigned
For each rule:
- Optionally add a condition of:
- Issue fields condition
- Field: Request Type:
- Condition: is one of
- Value: request types that you want to limit the rule to
- New action
- Notifications - Send web request
- Webhook URL:
<base URL>/<action> - (Optional) Headers
- HTTP method:
POST - Webhook body:
Issue data
- Webhook URL:
- Notifications - Send web request
For the webhook URLs, the action value should be one of the following according to the trigger:
createcommentorg-changetransitionassignment
For the JIRAHOOK capability, this must still be defined as a Jira webhook rather than as a Service Desk automation rule.
If using the framework with Zappa (see below), the optional headers can be used to set x-api-key to the appropriate API Gateway key value. This must be done on all or none of the webhooks.
On a development system, Flask is used to run the code.
Development of the code is done using a Pipenv-maintained virtual environment. The required modules are listed in Pipfile and can be installed with:
pipenv install
Note that Python 3 is required.
The additional modules required for development can be installed with:
pipenv install --dev
Unit tests and coverage tests can be executed with:
pipenv shell
pytest
If any Python packages are added via Pipenv, please remember to update requirements.txt as the latter is used by the Docker container's build process. Note that only the "top-level" packages are included, rather than generating the file automatically from Pipenv.
One way to troubleshoot the code is as follows:
- Modify the webhooks so that the content is delivered to RequestBin rather than the framework. This allows you to capture the data sent from Service Desk.
- Save the content to a file, e.g. "comment.json"
- Modify "local_test.py" to read the correct file and trigger the correct path.
- If using VS Code, use "Run and Debug" on "local_test.py". It is necessary to select "Raised Exceptions" in order to get VS Code to stop when the code breaks. Please note that there are some valid occurrences of raised exceptions in the framework.
Please note: if testing the comment route, please note that the framework explicitly reads the latest comment from the ticket, rather than using the comment data sent in the payload.
Contributions to the framework are more than welcome. Please ensure that tests are provided for all new code and that code coverage is maintained.