Designing confidence into a zero-downtime upgrade.
I designed the end-to-end administrator experience for Jira Data Center upgrades — from readiness and compatibility checks to monitoring a live upgrade and handling the moments when "zero downtime" could still contain downtime.
The system behind the screen
An upgrade wasn't a single action.
For a Jira administrator, upgrading a Data Center environment meant coordinating a sequence of readiness checks, compatibility decisions, system states and node-level progress.
The challenge wasn't simply helping someone click "upgrade." It was helping them understand what was happening well enough to act with confidence.
- 01Prepare
- 02Enter upgrade mode
- 03Update + monitor
- 04Finish
Readiness
Active operation
Intermediate states
Completion
Explanatory portfolio diagram — not Atlassian's technical architecture.
System modeling
Before designing screens, we had to model the operation.
The upgrade journey contained branching conditions, compatibility checks, node-level operations, warnings, decisions and exception paths.
I worked through the end-to-end behavior with the PM and developers, with my manager participating in some working sessions. The resulting system flow became a shared model for translating technical behavior into an administrator-facing experience.
Research
Would administrators trust themselves to upgrade?
I owned the research and design for the experience. We tested the workflow with Jira administrators, with the PM joining interview sessions. Beyond basic usability, I was interested in confidence: did administrators understand the process well enough to proceed with an upgrade without fearing they would get stuck or leave their Jira environment in an uncertain state?
I synthesized the sessions and produced the final research reports.
- 01
Do I know whether my environment is ready?
- 02
Do I understand what will happen when I start?
- 03
Can I tell when the upgrade is safe to finish?
Experience model
Three stages made the operation navigable.
- 01
Pre-upgrade
Establish readiness.
- 02
Upgrade
Start and monitor the operation.
- 03
Finish
Confirm completion and return to normal operation.
01 · Pre-upgrade
Establish confidence before anything changes.
Before entering the active upgrade, administrators needed visibility into the conditions that could affect the operation. The experience brought critical readiness information into the upgrade workflow, including instance health, add-on compatibility and Jira version readiness.
Compatibility
Make problems visible before they become upgrade problems.
The compatibility experience exposed individual add-on states against the target Jira version, allowing administrators to see what required attention before progressing.
Designing the exception
"Zero downtime" still had an exception.
Certain upgrades could require database index changes that introduced downtime. That created a communication problem as much as a technical one. The interface needed to make the exception visible, explain what the administrator was facing, and provide enough information to choose how to proceed.
The goal wasn't to hide the exception. It was to make the consequence understandable enough for an administrator to make an informed decision.
02 · Upgrade
Once the operation started, the interface became a monitoring system.
During the active upgrade, administrators needed more than a progress bar. They needed visibility into the state of the environment while individual parts of the system moved through the upgrade.
Progress wasn't just a percentage. It was system state.
Intermediate states
The hardest part was everything between start and done.
03 · Finish
Completion needed to be explicit.
The workflow closes with an explicit transition from active upgrade to completion rather than treating completion as an invisible backend event.
Iteration
The interaction model evolved with the technical details.
As we worked through the underlying behavior with engineering, the design evolved. The later pre-upgrade experience made readiness more explicit by bringing instance health, add-on compatibility and Jira version checks into a clearer model before administrators entered the active upgrade.
Ownership
My role
- Primary designer
- I owned the end-to-end UX and UI design.
- Research
- I planned and conducted administrator research, with the PM participating in interview sessions, and produced the final research reports.
- System modeling
- I worked with the PM and developers to map the branching upgrade behavior before translating it into the product experience.
- Collaboration
- I worked with a PM, Tech Lead, UX Lead and multiple developers throughout the project.
Outcome
The experience shipped in Jira Data Center.
The final experience shipped as part of Jira Data Center. Because reliable historical product metrics are not available, what can be stated is what was delivered:
A guided administrator experience spanning readiness, compatibility, active upgrade monitoring, exception handling and completion.
Shipped · Jira Data Center
Reflection
What stayed with me
Complex enterprise UX isn't always about hiding complexity. Sometimes confidence comes from exposing the right complexity at the right moment.
Zero Downtime Upgrades shaped a principle that has continued through my enterprise design work: when users are responsible for consequential operations, simplicity alone isn't enough. They need to understand the state of the system, the consequences of their choices, and what happens next.
Atlassian archive
More from Atlassian
JQL Search
Designing a more approachable way to construct powerful Jira queries.
Case study coming nextShared Filters & Dashboards
Designing information-management workflows for complex Jira environments.
Case study coming next