Design Pattern to update specific related Issues

[Ingo Wenke]

I'm working on an integration of 3 Jira projects. Each project represents a business process of our organization.

Those 3 processes are interacting very strong and relates to each other.

• Process 1: Customer Order - a JSM Process
• Process 2: Build to Order Process from Contract Management to engineering to shipping
• Process 3: Handover process - the process to keep the business partner who have to inspect and hand over the goods after shipping 

While our Build to order is a very detailed process about the overall engineering and shipping, the other two processes are on a more abstract level to keep our customer and a business partner informed. 

We have a triplet of

• one customer order issue
• one build to order issue
• one handover issue

and following ticket lifecycle:

• a customer raises an order
• the order ticket generates an build to order issue via automation
• when a business partner is assigned in a custom field at the build to order issue, a handover issue is generated, too

During workflow transitions and updates of issue fields, data is transferred via automation between those three projects (or specifically 3 issues)

Connecting issues in Jira is done by linking issues.

Due to it's nature of Jira and linking issue, each link can edited by users.

So, while generating a ticket, those links are set, but may be changed manually afterwards.

My current solution to setup a reliable connection of tickets (as an example):

• custom field "customerOrderKey" as text and read only
• customOrderKey is added on the issue screen of our build to order issue
• when the build to order ticket is generated via automation by an order issue, automation set the customOrderKey value to the key of the customer order issue.

So, the reference to the customer order issue is stored at the build to order ticket and can not be changed by user.

When I need to update the customer order ticket, du to progress at the build to order issue, automation can use the value of field customerOrderKey.

This is done by using the branch rule of automation.

To be discussed:

• Can this be implemented more efficient due to it's nature of updating just one specific issue, so that I can avoid the search query?
• Is there a possibility to avoid implementing custom fields as read only to "remember" the references, generated by an automation?

On a DB level, all what we are talking about is a referencial integrity of a set of records by (foreign) keys.