TechPrufer
Get started
Atlassian
Atlassian logoAtlassian · Feature trail

Teamwork Graph Connector Orchestration

4 posts · Jun 2026 – Aug 2026

What this trail covers

Atlassian's Teamwork Graph Connectors for Forge are generally available, enabling modeling of external data as Teamwork Graph entities. This facilitates ingestion of data from external SaaS products, internal platforms, or custom systems into Atlassian's Teamwork Graph, making it accessible across Rovo Search, Rovo Chat, Rovo Agents, and Atlassian Analytics. The platform handles connection configuration, data modeling, and recurring ingestion jobs. This capability is now integrated with Microsoft 365 Copilot and Teams via the Rovo agent, allowing users to search across Atlassian and Microsoft tools simultaneously, get contextually rich answers, and take action directly from Teams conversations. The Jira Cloud for Microsoft Teams app is enhanced with Rovo's intelligence to understand natural language, enabling users to trigger Jira workflows and assign work to agents without leaving chat.

All Atlassian trails
  1. Rovo for Microsoft 365: Turn conversations into Jira workThis post introduces the integration of Atlassian's Rovo agent with Microsoft 365 Copilot and Teams. Rovo, which leverages the Teamwork Graph, now provides cross-system context (Jira, Confluence, GitHub, Bitbucket, SharePoint, etc.) directly within Microsoft 365. Key new capabilities include searching across all connected tools simultaneously, providing holistic answers that connect work items and decisions, and enabling actions (like creating Jira work items or updating Confluence pages) directly from Teams conversations. The Jira Cloud for Microsoft Teams app is enhanced to understand natural language, allowing users to trigger Jira actions like status changes or work item creation/assignment using simple commands within Teams chats.
  2. Production-grade ingestion with orchestrationThis post details the implementation of production-grade orchestration for Forge Teamwork Graph Connectors. It introduces the `graph:connector` orchestration block, `scheduleOrUpdateTask` for root tasks, and `scheduleChildTask` for hierarchical data processing. Key contributions include the `taskRunner` function for dispatching and executing tasks, explicit failure classification (e.g., `UNAUTHORIZED`, `RATE_LIMITED`, `RETRYABLE_ERROR`), and the use of `persistentTaskMetadata` and KVS for state management. It also outlines best practices for idempotency, task granularity, and status reporting, and references SDK methods for scheduling and status updates.
  3. Bring any data into Atlassian system of workThis post introduces the architecture, module design, and connection lifecycle for Atlassian's Teamwork Graph Connectors for Forge. It details the `graph:connector` module, including its structure for defining object types, capabilities (replicatesPermissions, syncFidelity, supportsIncrementalSync), authentication providers, and data source configuration (`formConfiguration`, `validateConnection`, `onConnectionChange`). It emphasizes that the `onConnectionChange` hook is responsible for scheduling ingestion tasks, bridging the gap between configuration and the execution model. The post outlines a four-layer developer mental model: connection definition, ingestion model, execution model, and surfacing model. It also provides a code example for the `onConnectionChange` function.