The Atlassian Data Center Initiative

Objective

A large tech company with close to 60,000 employees had several on-prem installations of Jira Software and Confluence across the organization and had a business objective to consolidate as many of them as possible. This consolidation and migration to Jira Software and Confluence (Data Center) was critical because it would:

Having multiple instances that perform the exact same functions is a huge waste of money, resources and infrastructure. It can also be a nightmare to keep the instances upgraded along with all of the various apps that had been installed.

Atlassian licenses have to be installed per instance and cannot be shared across instances. Thus, the client was absorbing a significant amount of unnecessary costs and in many cases, lacked the ability to link tickets across projects or display them in certain Confluence spaces because the instances were not integrated and they were not working with a single source of truth.

The client was also paying for multiple licenses for the same apps/plugins, but was unaware of this because the instances were spread out over the enterprise and the duplicate costs were not easily traceable to the company’s financial resources. 

With multiple instances, teams could not integrate their data, tasks and communications when they needed to work with other business units. This led to a lot manual labor, duplication and ultimately tasks not being completed because they were lost somewhere in the middle of the chaos.

The client had archaic and forced high availability and redundancy setups that were fragile and often broke–especially when performing upgrades. The client also had issues trying to keep all of the various instances current and often passed on upgrading because other instances were not ready to be upgraded.

The Problem

The client had been a long time user of the Atlassian products and thus, had a ton of legacy information in the form of tickets and projects in Jira and pages and spaces in Confluence. The client also had years worth of user and group data in multiple User Directories (UDs) that were a mixture of internal and external UDs that needed to be accounted for during the consolidation and migration processes. All of the work needed to be done in a manner that had the least impact on the users.

The RenWare Way

Using our unique approach to solving consolidation and migration issues, we put together a plan that consisted of an (8) month schedule and performed the following activities to ensure a smooth final rollout for the client:

This effort required multiple test environments: one that replicated the clients existing environment with the server on-prem versions of Jira Software and Confluence and one for the Jira  Software and Confluence data center versions. Doing this allowed the our team take snapshots of the various instances and massage the data, configurations, users, and groups for a smoother merging of the instances.

We began with consolidating the Jira Software User Directories (UDs) and removing old users and groups that were not linked to any issues, pages, projects, and spaces in the Atlassian ecosystem. Care was taken not to delete users linked to any data as this could cause major issues. Once the UDs had been consolidated, we updated the Active Directory (AD) query that was being used in the newly established UD to ensure only Atlassian users were being pulled into the system and not the entire organization. This also ensured that the client was working with accurate licensing counts for the Atlassian products and the installed apps.

The instances had years worth of old Jira Software projects and many, many leftover schemes and associated content such as issue types, workflows, custom fields, and so on. This unused, legacy content was impacting performance–especially around certain JQL filters that were pulling more data than they should have been. Scrum and Kanban boards were also cleaned up as well as old dashboards.

A deep dive into the Confluence instances was conducted and a removal of old spaces was performed. When spaces have a lot of pages and those pages contain a lot of attachments, it impacts performance. By cleaning up Confluence, we were able to regain valuable efficiencies and instance resources that allowed Confluence to shine.

Having all of these instances working in isolation led to a lot of duplication which could also be found in the client paying several times for the same apps (aka plugins) in both Jira Software and Confluence. A thorough review by our team provided the client with a much needed accounting of monies being spent over the past (7) years.

Because each Jira Software and Confluence instance pair were operating in isolation, there was a complex maze of clutter that had to be unraveled. Our experts worked with the stakeholders of each instance pair and guided them towards better practices and created a set of standards that provided the necessary reusability that worked for all parties.

Once our team had created a solid, errorless strategy for consolidating all the instances for Jira Software and Confluence and the installed apps, we began working on a similar strategy for migrating the new, consolidated bundle to the data center platform for each product. This strategy included how to upgrade and maintain the platform and each node for each product.

The Results

Our delivery team provided the client with consolidated instances for both Jira Software and Confluence--each setup consisting of four nodes along with a runbook detailing how to quickly spin up new nodes and perform zero downtime upgrades. Our solution not only provided the necessary high-availability and redundancy the client wanted, but it also provided the ability to scale and onboard the remaining on-prem Jira Software and Confluence instances for a complete enterprise-level solution. The consolidation also allowed the client to recognize significant cost savings in the form of Jira Software and Confluence licenses, associated Atlassian Apps licenses, reduction in infrastructure, and the number of resources to maintain the new environment.

During the assessment, our experts notices a lack of integration with the client’s CI/CD toolchain, so we included this integration into the overall deliverable at no extra cost to the client. The goal was to ensure the client was receiving the full benefits from their Atlassian investment. Once integrated, our team provided guidance on

  • How to view build and release information on Jira issues.
  • How to implement Triggers on workflows to auto-transitions issues based on actions performed in the CI/CD tools.
  • How to properly use Jira’s fix versions/release features to bundle issues and generate release notes.

At RenWare, we practice what we preach, so all work for this engagement was setup in Jira and Confluence. We created a Jira project and integrated it with a Confluence space. We also created project-level and management-level dashboards and pages so our progress was easy to track and our client was always working with the most up to date information.

The client also gave us high marks for using our our unique methodology, which ensures there is minimal disruption to the client’s daily business operations and project activities.

VP, DevOps & Cloud

"RenWare was a valuable partner and contributor to the successful completion of our Atlassian Data Center migration and consolidation effort. Their ability to not only translate our business requirements into actionable tasks, but to also have the foresight to integrate the end-solution into our existing CI/CD environment was invaluable. If you're looking for an Atlassian partner that provides high-quality, cost-effective solutions, I'd highly recommend you contact RenWare."