Filter Activity Stream

Filter Activity Stream for Jira

See what changed in any saved filter or JQL query, on a dashboard or on its own page.

Filter Activity Stream is a Forge app for Jira Cloud. It adds two places to read the stream.

• A dashboard gadget that lists recent activity for the issues in a saved filter or a JQL query.
• A global page (Apps menu → Filter Activity Stream) with the same stream and a query form. It remembers each user's last query.

It shows five kinds of events.

Board re-ranking ("Rank" changes) is left out. It's drag-and-drop noise.

What you can do with it

Point it at any saved filter or JQL. Pick one of your starred filters, type a filter ID or an exact filter name, or write JQL. An ORDER BY in your JQL is ignored, because the stream is always newest first.

Then narrow it down.

• Pick a time window of the last 24 hours, or 3, 7, 14 or 30 days.
• Choose any mix of the five event types in the gadget settings, and change it on the fly above the list.
• Show only activity by up to 20 selected people.
• Group by issue (the most recently active issue first) or by day in the viewer's time zone, or don't group at all.
• Page through 10, 25 or 50 events at a time.

CSV export covers all loaded events, with UTC time, local time, issue, summary, event, actor, field, from, to and comment excerpt. The CSV opens in a dialog for you to copy into a spreadsheet, since UI Kit apps can't start a file download. Cells that start with =, +, -, @, tab or CR get a leading ', so spreadsheets don't run them as formulas. The export is limited to 2 MB.

When a limit is hit, a warning says which one. For example "Showing the newest 2,000 of 2,345 events."

Every Jira call runs as the person viewing the gadget. So restricted comments and issues they can't browse never appear, and the cache is per user.

Limits

How it works

Dashboard gadget / global page (UI Kit, render: native)
  └─ invoke('getActivity' | 'exportCsv' | 'listFilters' | 'getPageConfig' | 'savePageConfig')
       └─ resolver (src/index.js → src/lib/service.js)
            ├─ Forge KVS cache  (key = sha256(accountId + query); 5-minute TTL)
            └─ api.asUser().requestJira(...)
                 1. POST /rest/api/3/search/jql
                      jql  = (<filter or JQL>) AND updated >= -<window>m ORDER BY updated DESC
                      fields = summary, created, creator, reporter, updated [, comment]
                      expand = changelog          (only when change events are wanted)
                      pages of 100 via nextPageToken, at most 500 issues
                 2. GET  /rest/api/3/issue/{id}/changelog  (only for issues whose embedded changelog is partial;
                                                            Jira returns it oldest first, so the app reads the
                                                            last page first and walks back to the window start;
                                                            up to 100 issues, 3 pages of 100 each, 4 in parallel)
                 3. GET  /rest/api/3/issue/{id}/comment?orderBy=-created   (only when the embedded comments are
                                                            partial; 4 in parallel; stops at the window start)
            → events: dedupe by id, filter by window/type/person, sort newest first, cap at 2,000

Each event has a stable id, so duplicates collapse. Changelog items use h:<historyId>:<field>:<n>, comments use c:<commentId> and creation uses n:<issueId>. The same history or comment can come from the search and from a follow-up call, and you'll see it once.

The stream is newest first. When events share a timestamp, an issue's creation is listed below the edits made with it. After that, events are ordered by type (transition, assignment, field change, comment) and then by id, so the order is the same on every refresh.

Filter names are quoted and escaped as JQL strings (" → \", \ → \\), and numeric IDs are used as-is. Before your JQL is wrapped in ( … ), the app checks it for balanced parentheses and quotes. That check follows JQL's backslash escapes inside and outside quotes, and it refuses a backslash outside quotes that isn't followed by a character. So your JQL can't close the wrapper early. It always runs with your own Jira permissions, and paths sent to Jira come from a fixed allow-list (numeric issue IDs only).

Jira errors become plain messages. Bad JQL shows Jira's own explanation, and 403 or 404 says you lack permission or the filter is missing or not shared with you. The app retries 429 and 503 twice using Retry-After (at most 3 s per wait), within the time budget. If a follow-up call for change history or comments fails, even with a timeout or network error, the stream still shows with a warning. Forge's "allow access" (user consent) error passes through untouched, so Jira can show its consent prompt.

Permissions and data

The app has no external permissions, no remote back end and no egress. All code runs on Atlassian's Forge platform, so it's built to qualify for Runs on Atlassian. Atlassian decides eligibility, and you can check it with forge eligibility after deploying.

PRIVACY.md lists what is stored and for how long.

Known limitations

• CSV is copy-and-paste, not a file download. UI Kit has no download API.
• It hasn't run on a live Jira site yet. Unit tests with mocked Jira responses cover the logic, and the UI Kit screens were syntax- and import-checked but not rendered. NOTES.md has the smoke test to run after the first deploy.
• Comment edits and deletions aren't events. Only new comments show up, and a deleted comment disappears from the stream.
• Field changes show Jira's display strings, cut to 200 characters (long descriptions get shortened).
• There's no email digest in this version.

Support Zovo

Zovo builds and maintains this app. If it saves you time, you can support the work on GitHub Sponsors. Sponsorship is optional, and the tool works the same without it.

Zovo is made by theluckystrike on GitHub.

More tools by Zovo

• Dashboard Notes. Puts a formatted note on any Jira dashboard.
• Description Diff. Shows what changed in a Jira issue description, word by word.
• Epic Totals. Adds up story points, estimates and time logged for the children of a Jira epic.
• Expression Validators and Conditions. Adds workflow validators and conditions to Jira, with a test on a real issue.

License

MIT, see LICENSE.