As a technology-first company, Northwestern Mutual has been relocating their infrastructure to the cloud for years. When it came to Jira, the combination of increased user limits and data residency cleared the way, and the cost savings and native features in Atlassian Cloud provided clear advantages to the business.
Together with Expium—a Valiantys Company and Northwestern Mutual’s partner of record—the team set out to migrate one of the industry’s larger and more complex instances of Jira.
Northwestern Mutual’s environment details
How to justify the migration effort
Above all else, the effort to migrate Northwestern Mutual’s instance to the cloud had to make business sense.
Northwestern Mutual found a 25% savings in total cost of ownership when comparing Atlassian cloud to their AWS-hosted Data Center instance (understand total cost of ownership with Atlassian’s saving calculator).
They also opted for the premium tier which includes 24/7 Atlassian support with 1-hr response times for critical issues, and 99.9% uptime Service Level Agreement.
Then, Atlassian’s security model allowed Northwestern Mutual to proceed with confidence.
- Optimize & shift: Cleaning up the data prior to migrating it to cloud in a singular cutover.
- Lift & shift: Migrating the data as-is in a singular cutover.
- Phased migrations: Migrate the data in pieces, over a peri- od of time. Many times this is project by project.
Fresh start: Leave all your data behind, and start in a fresh instance.
Each strategy has its own optimal situation. For example, Fresh Start is commonly used for smaller instances that have less historical data, or where the admin team would like to take a new approach in the future instance.
It is worthwhile to be very familiar with the current state of the instance and understand the capacity for improvements after the migration. Northwestern Mutual opted for Optimize and shift, and used the migration as an opportunity to tidy up data prior to migration. The team archived data older than 36 months and modified the data to ensure it would transition smoothly to cloud.
- JCMA: Jira Cloud Migration Assistant is receiving heavy investment by Atlassian, and will be the preferred method for many future migrations.
- XML: A data export and import method that moves practically everything. It is much more manual than JCMA or CSV, and re- quires considerable data manipulation.
- CSV: Imports issue data, and requires the rebuilding of project structures and users in the destination environment.
Plan ahead—talk with your partner about the scope of your migration, and make sure to submit it for inclusion in your company’s budget cycle.
Because the Jira instance had vital data and a configuration that could not be migrated with JCMA, Northwestern Mutual opted for an XML migration.
Based on desired budget and risk appetite, the team decided to move Jira first, and not Jira and Confluence together. Jira Cloud was selected first because it provided more immediate benefits to the organization, and the users were minimally impacted by this approach.
Apps greatly increase migration complexity. It would be a mistake to overlook them when planning a migration. Many apps do not work in the same way in the cloud. It is imperative to investigate parity and functionality, and work out these differences with your users. In some cases cloud apps are bet- ter, in some cases they are just different.
App migration is not yet supported by Atlassian, which means that many apps do not move (yet) with JCMA
Expium created custom scripting to bridge the gap between current state and cloud for apps that do not have a clean migration path.
Additionally, apps are not currently required to follow Atlassian’s data residency settings. Each vendor has its own data residency and security standards. It may be necessary to contact a vendor directly to learn more. Investigate each cloud app’s data residency policy and ensure they are in line with your business requirements.
The best way to prepare for a migration is to migrate in a sandbox environment. Expium migrated XML exports from Northwestern Mutual’s production instance into multiple sandbox cloud instances and then methodically ran as many tests as necessary until the team produced reliable, flawless migrations.
For example, Expium started with an app and a test migration. If it failed, they wrote scripts to bridge the gap between current state and cloud, and tried again. Failing fast meant learning fast.
Based on what was learned in these cycles, Expium created precise step-by-step guides for the actual cutover. This approach prevents as many surprises as possible thus mitigating the risk of a failed migration.
A failed migration could mean a disruption in business continuity. It is always advisable to have a rollback plan so that if any problems occur, the team can revert to a state that ensures business continues as usual on Monday, no matter which environment is in production.
Transparency and alignment
Expium uses Atlassian tools Jira and Confluence to actively manage expectations and maintain alignment of the internal team and stakeholders. Confluence served as the real-time project dashboard with convenient links to relevant details for all stakeholders.
Expium documented the environment, findings, communications, and runbooks. Whenever timelines and expectations needed to be revisited, the team relied on Advanced Roadmaps to help convey the message.
For managing day-to-day tasks, the team coordinated and tracked with Jira. This reduced the need for meetings, kept the team focused, and provided effortless reporting.
Jira and Confluence helped the team manage internal and leadership communication. Without good end-user communication, regardless of how smooth the migration is, users will be disrupted, confused, and possibly angry.
Setting expectations for end users
Northwestern Mutual’s product owners created a communication plan for the migration that started three months before the cutover weekend, and included multiple targeted announcements to affected users.
The team at Northwestern Mutual also created multiple avenues for users to obtain information about the migration and resolve any questions via how-to videos, FAQs, and people available to address questions or concerns.
About the Authors
Bryan Darr is an Atlassian Certified Expert at Expium, a Valiantys Company and has been helping organizations of all sizes to work smarter. Bryan believes that establishing trust is fundamental to any relationship and advocates for open communication, transparency, and empathy. He enjoys working with the complexities of enterprise systems, as well as supporting businesses through change management and stakeholder engagement.
Michael Stewart was the driving force behind Northwestern Mutual’s
cloud migration. He has over 20 years in the industry and over 10 years administering Atlassian tools. Michael currently serves as a Senior Systems Engineer and is the primary administrator for Jira, Confluence, and Jira Align.