The Jira Cloud connector for Gemini Enterprise provides access to your Jira Cloud data, allowing you to interact with your projects, issues, and user information.
Supported Jira Cloud versions
The Jira Cloud connector supports version 2 of the Jira Cloud REST API.
Supported actions
When the Jira Cloud connector is enabled, end users can use natural language commands in Gemini Enterprise to perform the following actions.
| Action | Description |
|---|---|
| Create comment | Creates a new comment on a specified Jira issue to add feedback, updates, or notes. |
| Update comment | Edits the text of an existing comment. |
| Create issue | Creates a new issue in a Jira Cloud project. To create one, you need to provide a summary, the project ID, and the issue type ID. |
| Update issue | Updates an existing issue in Jira Cloud. You need to provide the issue ID to identify it, and you can modify various fields such as the summary, description, and assignee. |
In addition to these, additional read-only actions are available.
Required scopes
To enable Gemini Enterprise to perform search and actions using the Jira Cloud data store, you need the following scopes:
| Connection mode | Scope | Purpose |
|---|---|---|
| Federated search and actions | read:jira-user, read:jira-work |
Required to fetch and display search results and user details in real-time. |
write:jira-work |
Allows the assistant to modify work items (e.g., updating issue fields). | |
write:comment:jira |
Enables the assistant to post new comments or replies directly from the search interface. | |
write:issue:jira |
Necessary to create new issues or perform major edits to existing ones. | |
| Federated search only | read:jira-user, read:jira-work |
Required to search for and view issues, projects, and user profiles. |
| Data ingestion and actions | read:jira-work, read:user:jira, read:group:jira, read:avatar:jira |
Required for full-scale data ingestion and syncing user/group metadata. |
read:issue-security-level:jira, read:issue-security-scheme:jira |
Critical for access control; ensures the ingested data respects Jira's security boundaries. | |
read:audit-log:jira |
Used to track changes and perform incremental syncs to keep ingested data up to date. | |
write:jira-work |
Necessary to perform write actions on work items within the ingested environment. | |
write:comment:jira |
Required to add comments to issues using the assistant. | |
write:issue:jira |
Enables creating or modifying issues. | |
| Data ingestion | read:jira-work, read:user:jira, read:group:jira, read:avatar:jira |
Required to crawl and index Jira content and user relationships. |
read:issue-security-level:jira, read:issue-security-scheme:jira |
Required to map permissions so only authorized users can search the ingested data. | |
read:audit-log:jira |
Necessary for the sync engine to identify which items have been updated or deleted. |
For information on how to configure these scopes, see Configure minimum application scope.
Rate limits for data ingestion
The Jira Cloud connector supports a default rate limit of 12 queries per second (QPS) for data ingestion. These rate limits apply exclusively to the ingestion of data from the Jira Cloud connector.
Limitations
This section outlines known issues and limitations that may affect your use of the Jira Cloud connector.
Enforcing a VPC Service Controls perimeter on existing Jira Cloud data stores is not supported. To enforce VPC Service Controls, you must delete and recreate the data stores. For more information on VPC Service Controls and how to use actions after enabling VPC Service Controls, see Secure your app with VPC Service Controls.
The Jira Cloud data store is supported only in Global, US, and EU locations.
When creating a new application or adding a data store to an existing application, we recommend that you add a data store with actions belonging to a single connector type. For example, don't associate two Jira Cloud data stores with actions enabled to the same application.
The following are the limitations for the Jira Cloud ingestion connector:
For unstructured data, each media type has different restrictions. For more information, see Prepare data for custom data sources.
For structured data, each document must not exceed 500 KB in size.
Application roles with specific, non-empty roles cannot be granted
Browse projector security level permissions.Document permissions determined by user custom fields or group custom fields are not supported.
Jira Cloud doesn't allow any restrictions on worklog levels.
Jira Cloud doesn't index PDF attachments that exceed 2,000 pages.
Attachments added to private comments don't inherit the access restrictions (ACLs) on the comments.
Only the previous 20 comments for an issue in Jira Cloud are fetched.
Sprint tables can't be ingested.
The legacy user management model is not supported for integration with Jira Cloud. Only the centralized user management model is supported. For more information, see Atlassian Organization consolidation guide.
When querying a Jira entity (for example, an issue), related data like comments and attachments are not automatically included. This is because they are stored and indexed as separate entities. Searching based on the content of an attachment to find the parent issue is not supported.
What's next
- To create and configure a connector with Jira Cloud, see Set up a Jira Cloud data store.