Advanced Boards runs on Atlassian Forge and keeps its data in storage that Atlassian hosts for the app, kept apart for your site. It stores no text from your work items.

What the app reads from Jira #

The app reads, through Jira’s own interfaces:

• work items and their status history, to build each board’s flow;
• Jira boards, sprints and spaces, to seed a board and move work in sprints;
• the groups you belong to, to apply access policies;
• people’s names and avatars, saved filters, JQL, fields, statuses, link types, resolutions and the links between work items, to show people, check a board’s source as you, map a flow, match a backup and find blockers.

The app imports work items and their history in the background, as the app. Everything it shows you is limited to what you can read in Jira, checked as you.

Two kinds of scheduled work read Jira as the person who authorized them, never as the app: a forecast schedule, as the person it “Runs as”, and a risk rule’s evaluation, as the rule’s owner. Each reads only what that person can read, and stops when the authorization is gone. This is the only use of read:jira-work on someone’s behalf.

What the app stores, and what it never stores #

It stores identifiers, timestamps, status ids, the intervals it derives from them, the manifests of runs and evaluations, configuration, an audit trail and, for up to 30 days, the AI questions people ask with their answers. People are kept as Jira account ids. The app also keeps the names and avatar links Jira showed you for about an hour, until the next cleanup removes them.

It never stores text from a work item. Summaries, descriptions, comments and blocker reasons are read live from Jira, as you, each time a screen shows them.

How long each kind is kept is listed on Data retention and deletion. Most kinds are kept from 1 day to 360 days, and imported history for the window an administrator chooses (90, 180 or 360 days). Boards and their settings stay until they are deleted.

What the app sends out #

The app declares exactly two network addresses outside Atlassian, both Sentry ingest hosts for error reports: *.ingest.us.sentry.io and *.ingest.sentry.io.

• Error reports leave only a production installation, and only after a Jira administrator turns them on. Text from Jira, names and messages are removed first, and reports are kept for up to 90 days (Error reports).
• AI investigation sends its listed evidence and the question to an AI model through Atlassian’s Forge LLM service, and only after a Jira administrator turns it on (AI investigation settings). Work item summaries, comments, attachments, account ids, secrets, tokens and anything outside what you can read are never sent.

The app has no remote server and fetches no code at runtime.

The Jira permissions the app asks for #

Every change in Jira is made as the person who asked, and recorded in the audit trail.

Deleting your data #

A Jira administrator can remove the app’s data for the site with Delete all app data, after a preview and a typed confirmation. Jira work items, spaces and workflows are never touched.

Two things are kept:

• the counters of AI spend, so a deletion does not reset the AI allowance;
• one closing record of the deletion: who ran it, the reason and the counts.

Both are removed after about 90 days. The site’s monthly record of AI spend is never deleted, not even by an uninstall. To keep a board’s settings first, download its backup (Configuration backup and restore).