On this page
Short answer: plan for weeks, not days, for a large instance, and for a day or two per small project. Take an inventory, decide what you need history for, run a pilot on one project, import the rest in slices, validate with counts and spot checks, and keep Jira read-only for a while. The steps below work for any destination tool. The NUSK specifics are marked.
1. Take an inventory
List, for every project you will move:
- Issue types and how many of each
- Workflows and statuses
- Custom fields and who uses them
- Sprints, versions and boards
- Users, groups and permissions
- Attachments and how big they are
- Integrations (CI, Bitbucket, Slack) that point at Jira
Then cut the list. Archived projects and abandoned custom fields do not need to move.
2. Decide how much history to bring
You rarely need every change ever made. Choose:
- Open work: always.
- Closed work: how far back? One or two years is common.
- Comments, links and time logged: usually yes.
- Change history: nice to have for audit, usually not essential.
- Attachments: they make the move heavier. Bring them if people actually open them.
3. Map the model
Write a two-column table: Jira on the left, the new tool on the right. Cover:
- Issue types: epic, story, task, bug, sub-task and any custom types
- Statuses: which new status each Jira status becomes
- Custom fields: keep, merge or drop
- Sprints and fix versions: sprints and releases
- People: match them by email, so create the accounts first
4. Run a pilot
Pick one small, real project. Import it, then check issue counts per type, status counts, assignees, a few issues end to end with comments and links, and the sprint contents. Let the team use it for a real sprint. Only then plan the rest.
5. Import in slices
Importing everything at once makes problems hard to find. Import by project, and for large projects by date range. Choose a tool that skips issues it has already imported, so ranges can overlap safely, and that can roll an import back.
6. Validate
- Counts per project, type and status match the source, within your agreed exclusions.
- A sample of issues, about fifty across projects, match field by field.
- Links between issues and parent-child relationships survived.
- Saved filters and dashboards are rebuilt, because they usually do not migrate.
7. Cut over
- Pick a date. Announce it, with training and a short guide.
- Freeze Jira, or set it to read-only, on the day you import the final slice.
- Move integrations: commit message conventions, CI, notifications.
- Keep Jira read-only for a while for reference, then archive it.
How it works in NUSK
NUSK has a Jira migration built in, for Jira Cloud and for Jira Data Center or Server.
- Create users first. Assignees and reporters match by email, so create the NUSK users or turn on single sign-on.
- Connect. Open Settings > Jira migration, enter the site address and a token, and list the projects. Jira Cloud uses an account email and an API token. Data Center uses a personal access token, with the email left blank. The token is used for the import and not stored.
- Pick projects. Choose the projects and a delivery approach for each: agile, hybrid or predictive. NUSK creates one project per Jira project, and keeps the issue keys, such as
APP-123. - Import a range. In the project's import page, choose Jira, set Created from and Created to, and choose what to bring: attachments, watchers, work logs and change history. Leave the dates empty to take everything.
- Review the mapping. Map statuses (keep, map, or add as new), people, issue types and custom fields.
- Apply, and repeat for the next range. Issues already imported are skipped. Every import can be rolled back from the import history.
What maps to what
| Jira | NUSK |
|---|---|
| Project | Project, with a delivery approach you choose |
| Epic, Story, Task, Bug, Sub-task | The same work item types |
| Other issue types | Kept in a "Jira issue type" field |
| Status | Workflow status |
| Assignee, reporter, watchers | Owner, creator, watchers |
| Custom fields | Workspace fields |
| Sprints, fix versions | Sprints, releases |
| Comments, links, work logs | Comments, links, time entries |
| Change history | Revision history, read-only |
| Attachments | Project documents on the item |
Other routes. If you cannot give API access, import a Jira CSV export ("Export > CSV (all fields)"). Attachments are not copied from a CSV. Confluence pages can be imported into a project wiki by space key, keeping their tree, links, tables of contents and images.
Limits to expect. Jira returns only the latest comments on some very busy issues, and the import warns you. Change history is shown but cannot be reverted. Jira automation rules and marketplace apps do not migrate.
Running both tools side by side
For a few weeks, keep new work in one place only. Do not let teams work in both, or you will spend the savings on reconciliation. Use the old tool to look things up, not to create.
Common mistakes
- Importing before accounts exist, which leaves issues without owners.
- Migrating every custom field "just in case".
- Skipping the pilot.
- Forgetting the integrations that point at the old tool.
- No plan for people, only for data.
Start a Jira import · Jira alternative overview
See it with your own project
Import a plan or a Jira project into a free NUSK workspace and judge it on your own data.