Data residency scope

Last updated: 29.09.2026

Applies to: Visual Dependencies — Work Item Link Map for Jira.

Atlassian requires every partner that declares in-scope End-User Data in its app to publish what the app treats as in scope and out of scope for data residency. This page is that definition.

What “in scope” means here

Atlassian defines in-scope End-User Data only as “data that must comply with data residency requirements”, and leaves the definition to each partner. This is ours.

We treat data as in scope when it is the customer's own product data — the content of their Jira site and the records the app keeps on their behalf. That data stays inside Atlassian Forge hosted storage and follows the region the customer pins.

We treat data as out of scope when it is neither customer content nor attributable to a person from the data itself: measurements of how the software behaved, carrying an identifier we generated ourselves.

This is a narrower claim than “not personal data”. Some out-of-scope data is personal data under the GDPR and is treated as such — see our Privacy Policy. Our Legitimate Interests Assessment for product analytics is available on request.

In scope

Stored in Atlassian Forge hosted storage, pinned to the customer's region:

  • Saved filter presets, saved graph layouts (node positions), the last-used preset, and colour preferences — each keyed by an Atlassian account identifier.
  • The mapping from an Atlassian account identifier to the app's analytics identifier, the analytics opt-out record, and the record that the analytics notice has been seen.
  • Everything the app reads from Jira in order to draw a graph — work items, their links, assignees and reporters. The app reads this on demand and does not copy or retain it.

Out of scope

  • Product-analytics events. Content-free records of which app feature was used, how often, how long an action took and whether it succeeded, carrying the app's analytics identifier and coarse context: the surface the app was opened on, the licence edition, the Jira colour mode, a viewport size bucket, the environment and the app version. They contain no work-item content, keys or names, no project names, no Atlassian account identifier, no name or e-mail address, no page addresses, no browser, device or operating system, and no location. Processed by PostHog Inc. on PostHog Cloud EU.
  • The app's analytics identifier. A random value generated by the app per installation for each account. Never derived from the Atlassian account, name or e-mail address; never shared between customers; replaced after 12 months; deleted when the user opts out or their Atlassian account is closed.
  • Transport metadata of the analytics connection — the IP address the request arrives from. Inherent to any request leaving the browser; discarded by PostHog at ingestion, never stored by us, and never used for geolocation.
  • Atlassian's own operational and usage reporting about the app — install counts, active-user counts, invocation counts, error rates and latency. Produced by Atlassian as platform operator from data it already holds.

How the boundary is held

The out-of-scope claim rests on the payload, not on a promise, so it is enforced where payloads are built:

  • Every analytics property is a number, a boolean, or a value from a closed enumeration. There is no free-text property kind, so work-item content, keys, project names, filter text and user names are unrepresentable rather than merely forbidden.
  • A runtime allowlist drops any property not in that schema before a request leaves the browser.
  • Tests fail the build if the schema gains a free-text kind, or if person properties would be sent.

If the payload ever needed to carry customer content, the correct response is to declare that data in scope, with the consequences that has for how the app is certified — not to restate this definition.

Contact

Questions about this definition: [email protected]