What a lossless Jira import actually moves

Configuration first

Before a single work item moves, the wizard writes the configuration it read from your site: spaces and projects, work item types, the statuses that are actually used, custom fields with their options, versions, components and link types.

Then the parts other importers skip: workflows per type, roles with their people and permissions, boards with their columns, and Help Desk request forms with their customers and organisations.

Everything written in this step can be undone with one action, right up to the moment the first item lands.

Then the items

Descriptions are converted from Atlassian Document Format into GoProject rich text. Comments keep their internal flag. Attachments, work logs, watchers and participants, sprints, epics with their colours, parents, links, story points and start dates all come across. Board order is derived from Jira's rank.

Then again

Run the session a second time and only items changed since the last run are imported. Interrupt it and it continues from the last created item. Restart the server and stuck sessions are closed cleanly.

What we say no to

Workflow arrows from Jira Server and Data Center, because their API does not describe them: the statuses come across and you draw the arrows once in the workflow editor. SLA definitions from Jira Service Management, because the API does not return them. JQL filters, dashboards, automation rules and notification schemes, which are better rebuilt with GoProject's own tools than translated badly.

If you want to see it on your data, start free and import your own export.