The app runs no server and sends nothing anywhere. It reads your saved filters and counts how many work items each holds, inside your own Jira site.
Your settings, which are group names and the saved filters you put in them. These are saved in the gadget's own configuration, which Jira stores on your site.
To draw each row it asks Jira how many work items the filter matches. It receives a number. It does not store work items and does not copy them anywhere.
Every count runs as the person reading the dashboard. A filter that person cannot see shows a dash rather than a number, because an empty queue and a permission they do not have are not the same fact.
A count on its own does not say whether a queue is filling or draining. To answer that, the gadget keeps a short history per filter: the count and the time it was read, at most once every six hours and at most the last 120 readings. Nothing else is kept. No work item, no title, no name, no identifier of a person.
The history is stored in Forge storage, inside Atlassian, under your own site's installation of the app. It is not sent anywhere and we cannot read it.
Each reader's history is kept separately, keyed by a one way hash of their Atlassian account id, because the counts were made with that person's permissions. The account id itself is not stored, only the hash.
Removing the gadget or uninstalling the app removes the history with it.
Two scopes. read:jira-work, read only, nothing is written to Jira.
And storage:app, for the history described above, which lives inside
Atlassian.
None. No analytics, no telemetry, no advertising, no subprocessor. The app declares no external addresses, so it cannot make an outbound call even if it tried to.
Built on Atlassian Forge; it runs on Atlassian's infrastructure and we operate no server of our own for it.
If this policy changes, the date above is updated.