Jira on Cloud: Case Study for Migrating ITSM Solutions

• A runbook, which is basically a step-by-step list of tasks that Atlassian recommends for successfully completing the migration;
• A pre-migration checklist, designed to address many common issues up front that would prevent the Atlassian Server from communicating with the Atlassian Cloud;
• The Jira Cloud Migration Assistant (JCMA), which is a tool designed migrate project configurations and project data from the server environment to the Atlassian cloud.

It would seem, then, that the credit union, even with its limited IT staff, could pull off the Jira on cloud migration on its own. Unfortunately, it wasn’t that simple: it never is.

Even when SPK employed the very tools that Atlassian provides and recommends, there were issues. There were elements, for example, that refused to migrate.

Implementing Jira on cloud migration tools

For this project, SPK employed not only the migration tools described above, but also SQL and Jira administration expertise. After three weeks of project scoping, our Atlassian certified engineers, working in concert with our Computer Systems Analyst and others, quickly got deep into the weeds for the execution. We upgraded the memory and heap allocations on the old server, in order to allow JCMA to operate.

We also deep-dived into the data. For example, in order to have a successful migration to the cloud, every user must have a unique email address. This isn’t a requirement for the server-based version of Jira, so SPK needed to perform lots of detailed work, which included:

• Working with the customer to assess all of the 3rd party Jira application plugins, to determine what was needed in the cloud, and a strategy to migrate them outside of JCMA.
• Identifying duplicate accounts in Jira for the customer to resolve.
• Identifying accounts that did not have unique email addresses so the customer could then determine which of these accounts were required, and supply appropriate email address where needed.
• Setting up Atlassian access so that users in the cloud environment could access it with their global account and password.
• Using JCMA Dark Features to get around user IDs that did not necessarily need to be moved to the cloud.
• Identifying and backfilling tens of thousands of items that were missing required values that prevented the migration.
• Working with Atlassian support to resolve configuration issues that prevented JCMA from migrating information to the cloud.
• Performing multiple test migrations to iron out any further kinks in the migration process to produce an ironclad plan for the final migration.

Identifying problem areas within the migration

The server and Jira on cloud products are not identical. So, despite all the support tools and checklists, it does not provide a perfect one-to-one swap of technical features. JCMA is generally intelligent enough to identify potential issues which require post-migration adjustments, allowing the migration team to address each one.

But as noted above, one particular custom field on the server (which had been poorly configured when first created) could not be moved, preventing one of the client’s projects from migrating. In the end, the client opted to effectively leave it behind, in order to proceed with the overall migration.

Output and conclusion

In the short term, this credit union will no longer lose sleep at the prospect of losing its precious systems and data once Atlassian shuts off the lights. Longer-term, however, this SPK client will enjoy multiple benefits from the successful Jira on cloud migration:

• Security is improved.
• System updates are automatic: the credit union will always have, and be running, the latest version.
• Reliability is improved: the cloud-based infrastructure is far more robust.
• The platform is now more scalable: new Jira projects and other Atlassian products can be created with a simple click of a button.
• It’s more cost-efficient: since the new system is SaaS (software-as-a-service), the credit union only pays based on its registered number of users.