Microsoft announces official PagerDuty MCP server integration for Azure SRE Agent, enabling natural language incident management and on-call scheduling without local infrastructure.
Microsoft has announced the official PagerDuty MCP server integration for Azure SRE Agent, providing a seamless bridge between PagerDuty's incident management platform and Azure's AI-powered site reliability engineering capabilities. This cloud-hosted integration eliminates the need for local proxies, npm packages, or container deployments while enabling real-time interaction with PagerDuty's comprehensive incident management ecosystem through natural language.
What Changed
The integration introduces a hosted MCP server at mcp.pagerduty.com that connects Azure SRE Agent directly to PagerDuty accounts. The server supports both US and EU regional endpoints, with authentication handled through a single Authorization header using Token-based authentication. Since Azure doesn't offer a dedicated PagerDuty connector type, users configure the generic MCP server (User provided connector) option with manual authorization header setup.
Provider Comparison
Unlike traditional API integrations that require complex setup and ongoing maintenance, this MCP server approach offers several advantages:
- Zero infrastructure: No local proxies, containers, or npm packages required
- Real-time access: Direct connection to PagerDuty's hosted endpoint
- Natural language interface: AI-powered interaction with PagerDuty data
- Permission-aware: All actions respect the user account's PagerDuty permissions
- Comprehensive coverage: 60+ tools covering incidents, services, on-call, escalation policies, event orchestration, incident workflows, and status pages
Business Impact
For organizations using both Azure SRE Agent and PagerDuty, this integration significantly reduces the operational overhead of connecting these platforms. Teams can now leverage AI capabilities for incident investigation, on-call management, and service monitoring without the traditional integration complexity.
The ability to create specialized subagents with PagerDuty expertise means organizations can deploy focused AI agents that understand incident management workflows, escalation patterns, and on-call scheduling—all while maintaining proper access controls through PagerDuty's role-based permissions.
Key Capabilities
The integration provides comprehensive access to PagerDuty's platform through natural language commands:
Incident Management
- Create, list, and manage incidents with full lifecycle support
- Add notes and responders to ongoing incidents
- View and manage alerts associated with incidents
- Find related and historical incidents for investigation
- Identify outlier incidents based on patterns
On-Call & Scheduling
- List current on-call assignments across teams
- Manage schedules and create temporary overrides
- View schedule details and user assignments
- Handle escalation policies and routing
Service Management
- List and configure services with full CRUD operations
- Manage integrations and service settings
- Monitor service health and status
Event Orchestration & Workflows
- Configure event routing and processing rules
- Manage incident workflows and automation
- Handle status page communications during incidents
Implementation Steps
Step 1: Create PagerDuty API Token
Generate a User API Token from PagerDuty's User Settings > API Access section. The token inherits all permissions of the creating user account, so consider using a dedicated service account with restricted permissions for production deployments.
Step 2: Configure MCP Connector
In Azure SRE Agent, add a connector using the MCP server (User provided connector) option:
- Connection type: Streamable-HTTP
- URL: https://mcp.pagerduty.com/mcp (or EU endpoint for European accounts)
- Authentication: Custom headers with Authorization: Token
Step 3: Create Specialized Subagent (Optional)
Deploy a PagerDuty-focused subagent with custom system prompts and tool access. This subagent can be configured with PagerDuty expertise, incident investigation workflows, and best practices for on-call management.
Step 4: Add PagerDuty Skill (Optional)
Create a skill configuration that provides contextual knowledge about PagerDuty's platform, including troubleshooting guides, common workflows, and permission management.
Security Considerations
The integration maintains PagerDuty's security model through user-scoped permissions:
- Token-based authentication: Single User API Token in Authorization header
- Permission inheritance: Token permissions match the creating user's role
- Role-based access: Observer, Responder, Manager, and Admin roles control capabilities
- Audit logging: All API usage can be monitored through PagerDuty's audit logs
For production environments, Microsoft recommends creating dedicated service accounts with restricted roles to limit the blast radius of potential token compromise.
Troubleshooting
Common issues and solutions include:
- 401 Unauthorized: Verify token format (Token , not Bearer) and validity
- 403 Forbidden: Check user permissions and PagerDuty role assignments
- Connection refused: Ensure firewall allows HTTPS to mcp.pagerduty.com
- No data returned: Verify token user has access to requested resources
Available Tools
The integration exposes 60+ tools organized by function:
- Incidents: get_incident, list_incidents, create_incident, manage_incidents, add_note_to_incident, add_responders
- Services: get_service, list_services, create_service, update_service
- On-Call & Schedules: list_oncalls, get_schedule, list_schedules, create_schedule_override
- Teams & Users: get_team, list_teams, add_team_member, remove_team_member
- Event Orchestration: get_event_orchestration, append_event_orchestration_router_rule
- Incident Workflows: list_incident_workflows, start_incident_workflow
- Status Pages: create_status_page_post, create_status_page_post_update
- Log Entries: get_log_entry, list_log_entries
Regional Considerations
The integration supports both US and EU service regions:
- US (default): https://mcp.pagerduty.com/mcp
- EU: https://mcp.eu.pagerduty.com/mcp
Organizations must use the endpoint matching their PagerDuty account's service region to avoid authentication and connectivity issues.
Self-Hosted Option
While the hosted service is recommended for Azure SRE Agent integration, PagerDuty also offers a self-hosted MCP server. The self-hosted version only exposes read-only tools by default, with write operations requiring the --enable-write-tools flag at startup.
Conclusion
The official PagerDuty MCP server integration represents a significant advancement in how organizations can connect their incident management platforms with AI-powered SRE tools. By eliminating infrastructure complexity while maintaining security and comprehensive functionality, Microsoft and PagerDuty have created a solution that enables teams to focus on incident resolution rather than integration maintenance.
Organizations can now deploy AI agents with PagerDuty expertise, automate incident investigation workflows, and maintain proper access controls—all through a cloud-hosted integration that requires minimal setup and ongoing management.
For more information, visit the PagerDuty MCP Server documentation and Azure SRE Agent documentation.
Comments
Please log in or register to join the discussion