AWS Builder Center
AWS DevOps Agent - Atlassian Integration Setup

AWS DevOps Agent - Atlassian Integration Setup

Connect your AWS DevOps Agent with Atlassian's Jira and Confluence using the Rovo MCP Server. This integration enables your DevOps Agent to search Confluence documentation, retrieve page content, and create Jira tickets — all from within your agent workflows.

Prerequisites

  • An Atlassian Cloud instance with admin access
  • AWS DevOps Agent access
  • User permissions for the target Confluence spaces and Jira projects



Step 1: Trust the DevOps Agent Domain in Atlassian

Before the DevOps Agent can communicate with the Atlassian Rovo MCP Server, you need to allowlist the AWS domain.
  • In your Atlassian admin console, navigate to Administration → Apps → Platform experiences → AI Settings → Rovo MCP Server.
  • Add the following domain:  https://*.api.aws/** 
Atlassian Rovo MCP Server Settings




Step 2: Register the Atlassian MCP Server in DevOps Agent

  • In the AWS DevOps Agent console, go to Capability Providers → MCP Server.
  • Click Add MCP Server and configure the following:
Field
Value
Name
 Atlassian 
Endpoint URL
 https://mcp.atlassian.com/v1/mcp 
Enable Dynamic Client Registration
✅ Enabled
  • Click Next.
AWS DevOps Agent Add MCP Server

  • Select the apps to grant access to: Jira and Confluence.
Atlassian apps access

  • Review and confirm the access permissions for the selected apps.
Atlassian access confirmation

The Atlassian MCP Server should now appear as registered and ready for use.
> Note: Ensure that the user account associated with the DevOps Agent has access to the relevant Confluence spaces and Jira projects.



Step 3: Connect the MCP Server to an Agent Space

  • Open the Agent Space you want to integrate with Atlassian.
  • Click Capabilities.
  • Scroll down to the MCP Server section and add the Atlassian MCP server.
 Agent Space Capabilities - MCP Server

  • Select the tools you want to enable for this agent space.
⚠️ Important: Some tools perform write/edit operations (e.g., creating Confluence pages, editing Jira issues). Only enable the tools your workflow requires.
MCP tool selection list

  • Note the Webhook URL and Webhook Secret displayed after configuration. These are not required for this setup but may be useful for future integrations.



Step 4: Configure the Runbook Skill

Add a skill to your Operator Access configuration that instructs the agent on how to use the Atlassian tools during investigations.
  • Open the Operator Access web interface.
  • Click the Settings icon and select Add skill.
Add a new Skill

  • Provide instructions for the skill. Below is a recommended template:
## Investigation Protocol

### 1. Context Gathering (Do First)
Before beginning any investigation, search Confluence for relevant context:
- Use Atlassian/search with keywords related to the alert, service name,
or error message.
- If search returns a fragment, retrieve the full page using
Atlassian/getConfluencePage.
- Look for: runbooks, past incident postmortems, architecture diagrams,
service ownership, and on-call contacts.

### 2. Investigation
- Correlate Confluence context with live telemetry and logs.
- Reference past root causes from Confluence to avoid redundant investigation.
- If a runbook exists for this scenario, follow its steps before freelancing.

### 3. Documentation (Do Last)
After completing the investigation, create a Jira ticket using
Atlassian/createJiraIssue with:
- issueType: "Task"
- Summary: Brief description of the finding
- Description must include:
- What was detected (alert/signal that triggered the investigation)
- What was found (root cause or hypothesis)
- What was done (actions taken, commands run)
- What remains (recommended next steps for a human)
- Links to relevant Confluence pages consulted
- Severity assessment: Critical / High / Medium / Low
### Guardrails
- Do NOT create duplicate tickets. Search Jira first for existing tickets
related to the same issue before creating a new one.
- Do NOT modify existing Confluence pages unless explicitly instructed.
- If Confluence contains a runbook for the scenario, follow it.
- If no relevant documentation is found, note this in the ticket as a
gap — the team may need to create a runbook.



Tool Reference

Tool
Purpose
Operation Type
 Atlassian/search 
Search Confluence pages by keyword using natural language
Read
 Atlassian/getConfluencePage 
Retrieve the full contents of a specific Confluence page
Read
 Atlassian/createJiraIssue 
Create a Jira ticket (e.g., for investigation results)
Write



Verification

After completing the setup, verify that the integration is working:
  1. Open your Operator Access Web Interface for the configured Agent Space.
  1. Chat with DevOps Agent and ask questions related to content in Atlassian (e.g., "Search Confluence for our incident runbook").
  1. Ask the agent to create a test Jira ticket (e.g., "Create a Jira task titled 'Integration Test' in project XYZ").
  1. Confirm the Confluence content is returned and the Jira ticket is created successfully.



Troubleshooting

Issue
Resolution
MCP Server not connecting
Verify the domain  https://*.api.aws  is added in Atlassian AI Settings
Tools not appearing in Agent Space
Confirm the MCP Server is registered and the tools are selected in Capabilities
Permission errors on Confluence/Jira
Ensure the user account has access to the target spaces and projects
Search returns no results
Verify the Confluence space contains indexed content and the user has read access



Next Steps

  • Customize the Runbook Skill instructions to match your team's investigation workflows.
  • Add additional MCP tools as needed for your use cases.
  • Configure the Webhook URL/Secret if you plan to trigger agent workflows from external events in Confluence or Jira.
Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article