Feed The Troll - Jira Cloud App
Privacy & Security Policy
This policy describes how Feed The Troll handles data within your Jira Cloud instance. Effective date: 10 November 2026.
Introduction & Scope
This Privacy and Security Policy governs data processed by Feed The Troll (the “App”), a Jira Cloud application developed by drinkits and distributed via the Atlassian Marketplace.
This policy applies to all end-users, project members, and Jira System Administrators who interact with the App within their organisation’s Atlassian Jira Cloud instance.
What the App Does
Feed The Troll gamifies your team’s existing Jira workflow. When team members log time, transition issue statuses, leave comments, complete sprints, or send a teammate a Toast, the App awards XP to a personal “troll” character. Trolls level up, evolve through five stages, and collectively shape a shared team village. No new workflow is imposed. The App only looks at activity that already happens in Jira.
Each Team Village can also have a Village Guardian, a large creature in the village scene. Its shape comes from how the village’s Jira projects are set up, never from people or their work. See Village Guardian.
Data Storage (Forge Infrastructure)
App data is stored in two Atlassian services:
- Atlassian Forge SQL, a managed database. It holds the game: trolls, XP, quests, Toasts, villages and settings.
- Forge key-value storage (KVS). It holds a few small records: the state of the App’s own database setup, one admin preference, and the Village Guardian’s records. A short-lived Forge queue job starts the Guardian’s quick check; it carries only a random token.
Both run inside Atlassian’s infrastructure, and both are scoped to this App on your Jira site. There are no external servers, cloud buckets or off-platform databases.
What Is Stored and Why: Forge SQL
| Table | Contains | Retention |
|---|---|---|
trolls |
Troll profile per user per project: level, XP, lineage, cosmetics, privacy toggles, streak, last active date | Until app data is deleted |
xp_events |
Log of XP-awarding actions: event type, XP awarded, quality multiplier, source issue key, timestamp | Until app data is deleted |
daily_activity |
Daily XP and action count per user per project - used for the 30-day activity heatmap | Until app data is deleted |
kudos |
Sender ID, recipient ID, project ID, optional message (max 280 chars), XP awarded, source issue key | Until app data is deleted |
quest_progress |
Personal quest progress: quest ID, current count, target, completion status, timestamps, and what the quest has already counted (the days, or the issue keys) | Until app data is deleted |
inventory |
Cosmetic items earned via quests: item type, item ID, acquisition source | Until app data is deleted |
team_quests |
Team quest progress per project per sprint, and when an unfinished quest ended with its sprint - no personal data | Until app data is deleted |
villages |
Shared village state per project: building levels, defense score, prosperity score, and counts of work in progress and aging work - no personal data | Until app data is deleted |
raid_history |
Sprint raid outcomes per project: outcome, defense score, XP bonus, building affected - no personal data | Until app data is deleted |
raid_claims |
One row per project and sprint that has been raided, so a repeated sprint event cannot raid twice - project and sprint IDs only | Until app data is deleted |
project_settings |
Feature toggle configuration per project; stores the account ID of the last administrator who changed settings | Until app data is deleted |
global_settings |
Instance-wide settings (rollout strategy, Global Harmony Mode); stores admin account ID of last modifier | Until app data is deleted |
loot_boxes |
Tavern Loot Drops offered to a user in a project: tier, contents, opened state, timestamps | Until app data is deleted |
guilds |
Team Village name, configuration, and progression mode; stores the account ID of the admin who created or last changed it | Until deleted by an admin, or app data is deleted |
guild_project_bindings |
Which Jira projects belong to which Team Village - project and village IDs only | Until deleted by an admin, or app data is deleted |
troll_profiles, active_troll_profiles |
Traveling Troll profile per user per Team Village: level, XP, appearance, streak, and which profile is currently active | Until app data is deleted; archiving retires a profile without deleting it |
troll_profile_sources |
Which project trolls were merged into a Traveling Troll, so the merge can be explained back to the user | Until app data is deleted |
troll_profile_xp_events, troll_profile_quest_progress |
The Traveling Troll equivalents of the XP log and quest progress above | Until app data is deleted |
global_license_projects |
Which projects the free plan currently covers - project IDs only | Until changed, or app data is deleted |
product_metric_events |
Installation ID, activation metric name, and timestamp used only to prevent duplicate counter emission | Until app data is deleted |
troll_sprint_players |
The account IDs of players who turned on racing teammates in Troll Sprint | Until the player turns it off |
troll_sprint_runs |
For those players only: the UTC date of the daily track, the best time on it, that run’s ghost (the troll’s position 15 times a second), and whether that day’s Sprint Goal XP was paid | The track’s own day plus one day; deleted at once when the player turns racing off |
xp_award_slots |
Per player, per project or Traveling Troll, per kind of XP event and UTC day: how many awards were paid and when the last one was, so the daily caps and waiting times hold even when several Jira events arrive at once. Account ID, project or profile ID, event type, day and two numbers. | 3 days |
What Is Stored and Why: Forge KVS
| Record | Contains | Retention |
|---|---|---|
| Database setup | The state of the App’s database setup after an install or update: status, attempt and failure counts, timestamps and an error category. No personal data. | Until app data is deleted. The short claim records used while setup runs expire after 2 days. |
| Review invitation | One yes/no value: an admin chose not to see the Marketplace review invitation again. No account ID. | Until app data is deleted |
| Guardian readiness | Per Team Village: whether its projects could be read, how many were readable, and when the last reading ran | 14 days |
| Guardian candidate | Per Team Village: a one-way fingerprint of the projects’ structure categories, the Guardian’s root, three reason codes and timing | 3 days |
| Guardian reader state | Where the background reader stopped, a pause after Jira’s rate limit, and a one-minute mark while a reading runs. No project data. | 14 days and 2 hours |
| Guardian quick start | One per site: when a project change last started the Guardian’s quick check, a random token, and how many checks it has run. No project, village or account data. | 1 day |
| Guardian identity | Per Team Village, written once: the root, appearance values, rule versions, the UTC day it formed, the fingerprint and three reason codes | Until app data is deleted. It is never changed. |
| Siege Forecast counts | Per project: the four counts above (in progress, in progress for more than a week, flagged now, past due) and when they were taken | 14 days |
| Siege Forecast sprint | Per project: the running sprint’s ID, name and end date, for “the next raid” | 60 days, or until the sprint closes |
| Siege Forecast last raid | Per project: when the last raid ran, its sprint’s name, its outcome, its Defense and the factors that made it | 180 days |
| Jira wait | When Jira asks apps to slow down: until when the App waits before counting again. No project data. | The wait itself, at most 15 minutes |
A Guardian record names its village only by a one-way hash of the village’s ID, and none of the Guardian records holds a project ID, key or name, an account ID, issue data, XP or activity. The Siege Forecast records are kept per project under the project’s numeric ID, the same ID the SQL tables use; they hold no issue content, issue keys, people or comments.
On Uninstall
When the App is uninstalled from a Jira instance, the Forge platform stops serving the App's data and soft-deletes it. The data is then held for Atlassian's documented retention period before final deletion, so it is not necessarily destroyed at the moment of uninstall. Atlassian can relink or restore it if the App is reinstalled within 21 days. Feed The Troll runs no database of its own: nothing is stored outside Atlassian's Forge infrastructure, and the App cannot read the retained copy.
What Data We Process
The App processes the minimum data necessary to deliver the gamification experience. The following describes precisely what is read from Jira and what is derived and stored.
Data Read from Jira (Not Stored)
To detect XP-awarding events, the App receives event notifications from Jira and uses these parts of them, plus a few narrow reads for specific rules:
- Jira Account ID - the Atlassian-assigned opaque identifier for the user who performed the action (e.g.
5f7a3b...). This is not a name or email address. - Project ID - to associate the event with the correct village.
- Event type and timestamp - what happened (e.g. a worklog was created) and when.
- Issue key - the Jira issue identifier (e.g.
PROJ-42) stored alongside the XP event for the activity log. - Status category - whether an issue's new status is in Jira's To Do, In Progress or Done category, so a status counts whatever it is called.
- A sprint's projects and state - the projects of a sprint's issues (a JQL search that returns the project and nothing else), the board's project when a sprint has no issues, and whether a sprint is over, so its raid and team quests reach the right village.
- Counts of work in progress - per project, how many issues are in progress and how many of those have not moved for a week. These are JQL counts with no issue content, used for the village's defence.
- Counts for Siege Forecast - per project, also how many open issues are flagged now and how many are past their due date, and the running sprint's name and end date. Counts only, taken as the App.
- Blocking links, for Siege Forecast - when a person opens Siege Forecast, the App asks Jira, as that person, for the open issues in the village’s projects that have is blocked by links: each issue’s key, title, status and links. Jira applies that person’s permissions. The answer is shown to that person and is never stored or logged.
- Comment count - how many comments a flagged issue has, the number only, for the blocker bonus.
- Estimate and time spent - of an issue that was just finished, for the Estimation Sage quest.
- Project structure, for the Village Guardian, on sites where the Guardian is on. For each project in a Team Village, the App reads the project’s type and whether it is archived, its statuses per issue type, and whether it has any components and any released or unreleased versions. It reduces these at once to nine categories and keeps only the fingerprint described above. The answers themselves are never stored or logged.
Account IDs are shared between teammates on a team surface. To pick who to send a Toast to, the App needs to know who a teammate is, so the teammate list in a Team Village or project returns each teammate’s Atlassian account ID alongside their troll name and Jira display name. Only people who share at least one project the viewer can browse are listed, and the list never includes the viewer. The account ID addresses the Toast; the App does not use it to look up anything outside Jira, and it never leaves Atlassian’s cloud.
Jira API Scopes Requested
The App declares the following Forge permission scopes in its manifest. No other scopes are requested.
The App does not request write access to Jira issues, attachments, users, or any other Jira content. All write operations are limited to the App’s own Forge storage.
User-Controlled Visibility
Each user controls what teammates can see via five privacy toggles in the Gear tab of the My Troll issue panel. A save applies to every troll that person owns, not only the project in view. These settings do not affect data storage - they only govern what is displayed to other users within the App.
| Setting | Default | Governs visibility of |
|---|---|---|
| Show in Village | ON | Your troll appearing in the shared team village scene |
| Show Level | ON | Your level on the village leaderboard |
| Show Streak | OFF | Your consecutive active-day streak count |
| Show XP | OFF | Your total accumulated XP |
| Allow Toasts | ON | Whether others can send you a Toast (kudos) |
Admin-Controlled Visibility
A Jira System Administrator can also set Who sees trolls to Only your team. Then the troll is shown only to people who hold Jira’s View Development Tools permission, hiding it from anyone without that permission (such as external or portal customers) across the issue panel and the Team Village. This is a display control only - it does not change what data is stored. It is off by default.
Troll Sprint
Troll Sprint is a short running game in the Team Village. By default no time or result of a run leaves the browser or is kept in browser storage; the game only adds to two counts (the game opened, a run finished), described under Atlassian-Hosted Activation Metrics, and the day’s best time lasts only until the page is reloaded. The My Troll panel remembers in the browser only the date on which it last opened the game, so it invites a player once a day.
A player can turn on racing teammates in the game. It is off until that player turns it on, it is not offered under Harmony Mode or Ethics Mode, and the game shows exactly what is shared before it is turned on. For that player, the best time on each day’s track and its ghost are stored (see the table above). Teammates who can see the player’s troll in a village can race that ghost and see the troll name and the difference in time. The App carries the player’s Atlassian account ID in the race payload so a teammate can address a Toast to them, but it is not displayed in the Feed The Troll UI. Meeting the day’s Sprint Goal earns 10 XP once a UTC day, recorded like any other XP, so teammates see it only if the player shares their XP. A player who shares also sees on the Troll Sprint sign how many teammates ran today’s track, a number and never who. Turning racing off deletes the player’s runs at once. A Jira System Administrator can turn Troll Sprint off for the whole site.
Village Guardian
Each Team Village can have one Guardian. It is on for every site.
How it forms
In the background, the App reads the structure of the village’s projects (see Data Read from Jira). When two complete readings taken at least five minutes apart agree, the Guardian forms once and its identity is written. When someone saves a change to a village’s projects, the App runs those readings a few minutes later instead of waiting for the next twice-daily check, so a new village’s Guardian usually forms within about ten minutes. The quick check reads the same project structure and nothing more. Adding or removing projects later does not change it.
The Guardian is built only from project structure. Issue content, comments, worklogs, XP, levels, streaks and anything about the people in the village play no part, and none of them can change it. It is the same for everyone in the village and is not a score or a ranking.
What reaches the browser
The village page receives only what it needs to draw the Guardian: its root and appearance values. The fingerprint and the reason codes stay on the server.
The first time a person sees a Guardian wake, their browser remembers it in local storage, so the wake plays once per browser. The entry holds the value 1 under a short hash of the village’s ID. It is never sent anywhere. Clearing site data simply lets the wake play again.
Who sees it
Anyone who can open the Team Village and browse at least one of its projects. Who sees trolls applies to the Guardian too.
GDPR & Data Rights
For organisations with users in the European Union or European Economic Area, this section describes the GDPR framework governing the App.
Controller and Processor Roles
Your organisation (the Jira Cloud customer) is the Data Controller: you determine the purpose and means of processing personal data by choosing to install and configure the App within your Jira instance.
drinkits acts as a Data Processor operating through Atlassian’s Forge platform. We process only the data necessary to deliver the App’s stated functionality, in accordance with your instructions as expressed through the App’s configuration.
Individual Data Rights
The following rights are available to data subjects. Because the App stores only Atlassian Account IDs (not names or email addresses), identity verification is handled via Atlassian’s existing authentication.
■ Right to Access
View your troll profile, XP history, quest progress, and activity heatmap directly in the App, from the My Troll issue panel. Contact us for a complete data export.
■ Right to Erasure
Atlassian’s standard User Deletion APIs propagate to Forge SQL, removing your records when your Atlassian account is deleted. You may also contact us directly to request erasure of your troll data only.
■ Right to Portability
Contact us via the support portal to request a structured export of your personal troll data in a machine-readable format.
■ Right to Restrict
Use the five privacy toggles in the Gear tab of the My Troll issue panel to restrict the visibility of your data to teammates. For full processing restriction, contact us or ask your Jira System Administrator to remove your troll profile.
Data Retention
Data is kept while the App is installed in your Jira site. A few records expire on their own:
- Troll Sprint runs, the day after their track’s day, or at once when the player turns racing off;
- the Guardian’s readiness, candidate, reader and quick-start records, after 10 minutes to 14 days;
- Siege Forecast’s counts after 14 days, its sprint record after 60 days or when the sprint closes, its last-raid record after 180 days;
- the XP award counters behind the daily caps, after 3 days;
- the short claim records used during database setup, after 2 days.
On uninstall, the Atlassian Forge platform soft-deletes the App’s Forge SQL and Forge KVS data and holds it for its documented retention period before final deletion. You don’t need to do anything. Atlassian can relink or restore that data if the App is reinstalled within 21 days.
International Transfers
The App does not transfer data across borders independently. Any geographic placement of data is determined by Atlassian’s infrastructure and your organisation’s Jira Cloud region selection, and is governed by Atlassian’s Data Processing Addendum and Standard Contractual Clauses where applicable.
CCPA (California)
We do not sell personal information as defined under the California Consumer Privacy Act. California residents have the right to know what data is collected (this policy), to request deletion (contact us or uninstall the App), and to opt out of sale (not applicable - no sale occurs).
Contact Us
For any privacy or security question, request, or concern related to Feed The Troll, contact us through one of the following channels:
| Channel | Details |
|---|---|
| [email protected] | |
| Support Portal | drinkits.atlassian.net/servicedesk - preferred for formal data subject requests |
| Response time | Target initial response within 1 business day; formal data subject requests are completed within the period required by applicable law |
| Security disclosures | Mark your support ticket or email as Security Issue. Critical reports are triaged within 24 hours, contained as quickly as possible, and followed by written status updates. |
Policy Updates
We will update this policy when the App’s data practices change. Material changes will be announced via the Atlassian Marketplace changelog. The effective date at the top of this page will reflect the most recent revision.