Connect PagerDuty to AI: Give Claude, Cursor, and Windsurf Access to Your Incident Data

Why Connect PagerDuty to Your AI Agent?

During incidents, every context switch costs time. PagerDuty holds the critical data: who's on call, what alerts fired, incident timelines, escalation policies. Checking that information means opening the PagerDuty app, navigating services, filtering by time window, reading through alert details.

Connecting PagerDuty to your AI agent via MCP lets you query incident state in natural language without leaving your workflow.

What You Can Do

With PagerDuty connected to Claude, Cursor, or Windsurf:

• Ask "are there any active P1 incidents right now?" and get an immediate answer
• Say "who's on call for the payments team this weekend?" without checking the schedule
• Query "what alerts fired for the auth service in the last 6 hours?" during debugging
• Get an incident summary for a postmortem without manually compiling timeline events
• Check if an alert was already acknowledged before escalating

How to Connect (3 Steps)

1. Create the MCP server — Open DataFaucet, enter https://app.pagerduty.com, and log in. Browse incidents, check on-call schedules, and review services. Each action becomes a tool.

1. Deploy — Click deploy. Your PagerDuty MCP server goes live with a unique endpoint and API key.

1. Connect to your AI client — Add the server config to Claude Desktop, Cursor, or Windsurf. The agent can now query PagerDuty directly.

Supported AI Clients

Any MCP-compatible client works. Standard Model Context Protocol.

Security Considerations

• Use read-only PagerDuty roles for querying incident data
• Scope API tokens to specific teams rather than full account access
• All API calls logged in PagerDuty audit trail
• Rotate MCP server API keys on the same cadence as other credentials
• Consider separate servers for different escalation tiers

Real Workflow Example

Before: Open PagerDuty → navigate to Incidents → filter by service → click each incident → read timeline → check escalation policy → return to terminal.

After: "Summarize active incidents for the payments team, who's currently assigned, and have any been open longer than 30 minutes?" Answer in seconds.

Pairs Well With

• Slack MCP server — coordinate incident response alongside PagerDuty status
• Argo CD MCP server — correlate deployment failures with new incidents
• Datadog MCP server — check monitoring alongside alerting for full incident picture

Get Started

Connect PagerDuty to your AI agent in under 60 seconds. Browse app.pagerduty.com through DataFaucet, deploy, and start querying incident data from your editor.

Connect PagerDuty to AI →