Sumtree documentation (public KB source)
Get started (5 minutes)
- Create a field. Jira settings > Work items > Fields > Create field > Sumtree rollup. Give it a name people will search for, such as "Total story points".
- Choose where it applies. In the field's context, pick the projects and work types (for example Epic and Initiative). The field only exists, and is only calculated, there.
- Choose what it adds up. On the field's Contexts page, open the Sumtree configuration. Pick a preset or set it yourself; the sentence at the top always says what the field will hold. Jira's Save stores the configuration on the context.
- Fill existing work items. Jira settings > Apps > Sumtree > Recalculate. Leave the JQL
empty for the whole site or narrow it (
project in (WEB, MOB)). - Add it where people look. The field is read-only and shows in the issue view's details. Add it as a column in list view or search, and use it in JQL and filters.
From then on Sumtree updates values by itself.
What can a rollup field hold?
| Setting | Options |
|---|---|
| Value from each work item | Story points ("Story point estimate" or "Story Points", first one set), original estimate, time spent, remaining estimate, any number field, or work items (each counts as 1) |
| Combine as | Sum, average, minimum, maximum, count with a value, count without a value, percent done |
| Which work items | All levels below (subtasks, stories, epics, and any level above epics), or direct children only; optionally include the work item's own value |
| Status | Any, not done only, done only (by status category) |
| Work types | Only these, or all except these |
| Time | Minutes, hours, days or weeks, with your hours per day and days per week |
| Compare with a field on the work item | Total as % of a field ("Budget used %"), or field minus total ("Budget left") |
| Display | Decimals (0-4), a unit after the number (pts, h, %), empty or 0 when there is nothing to add up |
| History | Off by default: changes are not written to the work item's history |
Percent done is the share of the total that sits in Done-category statuses: by story points if the value is story points, by count if the value is work items. Not-done filters do not apply to it (it needs both sides).
All levels below adds each work item's own value once. A story's own points and its subtasks' points are separate values, so nothing is counted twice, and Sumtree never reads one rollup field to compute another.
When do values update?
- When anything below changes: a number the field reads, a status, a work type, a worklog, a new child, a child moved to another parent (both parents update), a deletion. Usually within seconds.
- When someone opens the work item: the value is computed fresh and saved.
- Daily: everything updated in the last day is rechecked, with all its parents (turn off in Settings).
- On demand: Recalculate on the admin page, or Recalculate now in the Rollup breakdown panel.
The Rollup breakdown panel
Open it from the issue view's right-hand panels (Rollups). For every Sumtree field that applies to this work item it shows the value, the configuration sentence, how many items were counted, the split by To do, In progress and Done, and each direct child with its share (including everything below that child). It lists only work items you can see.
Limits
- Largest tree: 5,000 work items below one top item by default, up to 20,000 (Settings). Larger trees are skipped and logged, never summed partially.
- Values are numbers. Jira shows them with the unit you set in list view, search and exports without the unit (so CSV round trips cleanly).
- Jira Automation may not trigger on app fields yet (an Atlassian limitation for Forge custom fields). Scheduled rules with JQL on the field work.
- Rollup fields are read-only, and like other read-only app fields they do not show on create or transition screens.
Uninstalling
Fields of the Sumtree type disappear when the app is uninstalled and come back with their values if it is reinstalled within 30 days (Atlassian's rule for Forge fields).
Privacy and security
Runs on Atlassian. No data leaves Atlassian. See the privacy policy.