← Back to directory

PagerDuty MCP Server

Official

Archived

PagerDuty's official MCP server for managing incidents and on-call from your IDE or chat client

How to install ↓ Compare GitHub source repository ↗

Category: Dev Tools #60 of 438
Stars: ★ 79 Popular
Transport: stdio (local process) · Streamable HTTP · SSE
Runtime: Python · Docker
Credentials: API key / credential required
License: Apache-2.0
Last commit: 37 days ago
Tools: 82

64FMRS · C

An officially maintained PagerDuty MCP server covering incidents, services, schedules, escalation policies, orchestrations, teams, status pages, and webhooks, read-only by default with write access opt-in. Comprehensive functionality, but write mode and the user-scoped API key carry real permission risk, and HTTP transports lack built-in authentication, so deployment needs care.

Strongest · Documentation 17/20 Weakest · Reliability 8/20

Reliability

8/20

Security and permissions

14/20

Maintenance

11/20

Documentation

17/20

Setup experience

14/20

Why each score

Reliability 8/20: The manifest declares a stdio transport via uvx and a comprehensive tool list with clear read/write annotations that is consistent with the README's tool table — no visible contradictions between declared and described behavior. However, no CI workflow files or test code are present in the reviewed material, so the MCP init/tool-list handshake and error-handling paths cannot be verified as reproducible. Per the static calibration rule, this caps the score at or below 12. The lack of any executable evidence for error handling or dependency failure behavior further limits the score.
Security and permissions 14/20: The project is fairly transparent about its main risks: only read-only tools are exposed by default, write operations require an explicit --enable-write-tools flag, and the README explicitly warns that HTTP transports (streamable-http/sse) have no built-in authentication, defaults to 127.0.0.1, and should only be exposed on 0.0.0.0 behind an authenticating proxy. It also flags the plaintext-credential risk in Claude Desktop's config file. No hardcoded secrets, real tokens, or red-line issues were found. Deductions: the server uses a full-privilege PagerDuty User API Token with no scoping/least-privilege mechanism; destructive or high-impact actions (delete_team, delete_webhook_subscription, manage_incidents) rely only on the global write-tools flag with no per-action confirmation; and the embedded React UI apps add attack surface that isn't documented in terms of data flow or sandboxing.
Maintenance 11/20: The repository is Apache-2.0 licensed, not archived, and at manifest version 0.2.1, indicating an early-to-mid stage project. No commit history, release cadence, or dependency-update evidence is provided in the given material. 35 open issues against 75 stars is a notable ratio but responsiveness cannot be assessed from this data. Being an official vendor (PagerDuty) project gives clear ownership, which is a positive, but the absence of concrete release/security-response channel evidence keeps the score in the mid-range.
Documentation 17/20: The README is layered and thorough: prerequisites, multiple client integrations (Cursor, VS Code, Claude Desktop), Docker deployment, a transport-mode comparison table, an environment variable configuration table, and a full categorized tool list marking read-only vs. write tools, plus a dedicated security section on HTTP transport exposure. These are all reviewable and consistent. Gaps: no concrete tool call parameter/response examples, no disclosure of PagerDuty API rate limits, quotas, or cost implications, and troubleshooting guidance is limited to basics like restarting the client.
Setup experience 14/20: The setup path is clear: install via uvx/pip, set one required environment variable (PAGERDUTY_USER_API_KEY), with concrete config examples given for Cursor, VS Code, Claude Desktop, and Docker — few steps, broad platform coverage. Per the static calibration rule, since no CI or test files verify these configurations actually work end-to-end, the score is capped below 15. Some manual care is also needed since the write-tools flag is documented via an inline JSON comment that isn't valid JSON, which could trip up users copying the config directly.

Static review · not runListed 2026-08-26

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkChanges third-party account data

Best for

SREs / on-call engineers already using PagerDuty
DevOps teams wanting AI-assisted incident triage without leaving their IDE
Teams that want to query incident metrics or manage schedules conversationally

Not for

Teams that don't use PagerDuty
Organizations requiring strict least-privilege, role-scoped tokens (this server uses a user-level API token with that user's full account privileges)
Deployments that expose the HTTP endpoint directly on untrusted networks without an authenticating reverse proxy

Required permissions

Requires a PagerDuty User API Token (`PAGERDUTY_USER_API_KEY`) with that user's full privileges on the PagerDuty account
Only read-only tools are exposed by default; `--enable-write-tools` must be explicitly set to create, update, or delete incidents, services, teams, schedules, escalation policies, webhooks, and status page content
Optional `PAGERDUTY_API_HOST` to target the standard or EU regional API endpoint

Risks and side effects

The streamable-http and sse transports have no built-in authentication — any client able to reach the host/port can invoke tools, including write tools, using the single API key the server was started with; the README explicitly warns to keep the default loopback bind and only expose 0.0.0.0 behind an authenticating reverse proxy/gateway
Write tools are powerful: they can delete teams, modify escalation policies and schedules, and create status page incident posts, so mistakes can disrupt real on-call routing or customer-visible status pages
Creating a webhook subscription returns a signing secret that must be stored securely
Claude Desktop's config file requires the API key to be stored in plaintext, so file permissions matter

Setup

Before you start

Runtime：Python · Docker

PAGERDUTY_USER_API_KEY requiredsecret Required PagerDuty User API Token, generated under User Settings > API Access in your PagerDuty account.

Other optional settings (3)

PAGERDUTY_API_HOST optional Optional API base URL; defaults to https://api.pagerduty.com, use https://api.eu.pagerduty.com for EU accounts.

MCP_HOST optional Optional listen address for HTTP transports, defaults to 127.0.0.1.

MCP_PORT optional Optional listen port for HTTP transports, defaults to 8000; must be a valid integer.

Generate a PagerDuty User API Token under User Settings > API Access.
Make sure uv/uvx is installed globally (or use a local checkout with uv run, or the Docker image).
Add a pagerduty-mcp server entry to your Cursor, VS Code, or Claude Desktop MCP config using uvx pagerduty-mcp, appending --enable-write-tools if you need write access.
Set PAGERDUTY_USER_API_KEY (required) and optionally PAGERDUTY_API_HOST (use https://api.eu.pagerduty.com for EU accounts) via environment variables.
Restart the client and confirm the connection by asking something like "Show me my latest incidents."

claude_desktop_config.json

{"mcpServers":{"pagerduty-mcp":{"command":"uvx","args":["pagerduty-mcp","--enable-write-tools"],"env":{"PAGERDUTY_USER_API_KEY":"your-pagerduty-api-key-here","PAGERDUTY_API_HOST":"https://api.pagerduty.com"}}}}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json

{
  "servers": {
    "pagerduty-mcp": {
      "command": "uvx",
      "args": [
        "pagerduty-mcp",
        "--enable-write-tools"
      ],
      "env": {
        "PAGERDUTY_USER_API_KEY": "your-pagerduty-api-key-here",
        "PAGERDUTY_API_HOST": "https://api.pagerduty.com"
      }
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal

claude mcp add pagerduty-mcp -e PAGERDUTY_USER_API_KEY=your-pagerduty-api-key-here -e PAGERDUTY_API_HOST=https://api.pagerduty.com -- uvx pagerduty-mcp --enable-write-tools

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

The client's tool list should include tools like list_incidents; sending a prompt such as 'Show me the latest incident' and getting back PagerDuty data confirms the connection works.

Troubleshooting

Verify `uv`/`uvx` is installed globally and reachable by the client (restart the terminal or IDE if needed)
Confirm `PAGERDUTY_USER_API_KEY` is set correctly; for EU accounts set `PAGERDUTY_API_HOST` to `https://api.eu.pagerduty.com`
If write operations (e.g., creating incidents, schedule overrides) fail silently, check that `--enable-write-tools` was passed at startup
In VS Code, use the command palette's `MCP: List Servers` to check server status and restart it if needed
When using an HTTP transport, verify `MCP_HOST`/`MCP_PORT` or `--host`/`--port` are set correctly and that 0.0.0.0 is not exposed directly to untrusted networks

Things to try

Once connected, you can ask your AI assistant things like:

Show me my latest PagerDuty incidents
List my event orchestrations
Show me the on-call manager
Add a note to an incident

Tools 82

create_alert_grouping_setting writes

Creates a new alert grouping setting

delete_alert_grouping_setting destructive

Deletes an alert grouping setting

get_alert_grouping_setting read-only

Retrieves a specific alert grouping setting

list_alert_grouping_settings read-only

Lists alert grouping settings with filtering

update_alert_grouping_setting writes

Updates an existing alert grouping setting

get_incident_metrics_all read-only

Retrieves aggregated incident metrics across all services

get_incident_metrics_by_service read-only

Retrieves incident metrics grouped by service

get_incident_metrics_by_team read-only

Retrieves incident metrics grouped by team

Show 74 more tools

get_responder_load_metrics read-only

Retrieves responder load metrics

get_responder_metrics read-only

Retrieves responder performance metrics

get_business_service_dependencies read-only

Retrieves dependencies for a specific business service

get_technical_service_dependencies read-only

Retrieves technical service dependencies for a business service

list_business_services read-only

Lists business services

get_change_event read-only

Retrieves a specific change event

list_change_events read-only

Lists change events with optional filtering

list_incident_change_events read-only

Lists change events related to a specific incident

list_service_change_events read-only

Lists change events for a specific service

get_event_orchestration read-only

Retrieves a specific event orchestration

get_event_orchestration_global read-only

Gets the global orchestration configuration for an event orchestration

get_event_orchestration_router read-only

Gets the router configuration for an event orchestration

get_event_orchestration_service read-only

Gets the service orchestration configuration for a specific service

list_event_orchestrations read-only

Lists event orchestrations with optional filtering

update_event_orchestration_router writes

Updates the router configuration for an event orchestration

append_event_orchestration_router_rule writes

Adds a new routing rule to an event orchestration router

create_escalation_policy writes

Creates a new escalation policy

get_escalation_policy read-only

Retrieves a specific escalation policy

list_escalation_policies read-only

Lists escalation policies

update_escalation_policy writes

Updates an existing escalation policy

add_note_to_incident writes

Adds a note to an incident

add_responders writes

Adds responders to an incident

create_incident writes

Creates a new incident

get_alert_from_incident read-only

Retrieves a specific alert from an incident

get_incident read-only

Retrieves a specific incident

get_outlier_incident read-only

Retrieves outlier incident information for a specific incident

get_past_incidents read-only

Retrieves past incidents related to a specific incident

get_related_incidents read-only

Retrieves related incidents for a specific incident

list_alerts_from_incident read-only

Lists all alerts for a specific incident with pagination

list_incident_notes read-only

Lists all notes for a specific incident

list_incidents read-only

Lists incidents

manage_incidents writes

Updates status, urgency, assignment, or escalation level

get_incident_workflow read-only

Retrieves a specific incident workflow

list_incident_workflows read-only

Lists incident workflows with optional filtering

start_incident_workflow writes

Starts a workflow instance for an incident

get_log_entry read-only

Retrieves a specific log entry by ID

list_log_entries read-only

Lists all log entries across the account with time filtering

add_team_member writes

Adds a user to a team with a specific role

create_team writes

Creates a new team

delete_team destructive

Deletes a team

get_team read-only

Retrieves a specific team

list_team_members read-only

Lists members of a team

list_teams read-only

Lists teams

remove_team_member writes

Removes a user from a team

update_team writes

Updates an existing team

get_user_data read-only

Gets the current user's data

list_users read-only

Lists users in the PagerDuty account

list_oncalls read-only

Lists on-call schedules

list_priorities read-only

Lists priorities defined in the account

create_schedule_override writes

Creates an override for a schedule

get_schedule read-only

Retrieves a specific schedule

list_schedule_users read-only

Lists users in a schedule

list_schedules read-only

Lists schedules

create_schedule writes

Creates a new on-call schedule

update_schedule writes

Updates an existing schedule

create_service writes

Creates a new service

get_service read-only

Retrieves a specific service

list_services read-only

Lists services

update_service writes

Updates an existing service

create_status_page_post writes

Creates a new post (incident or maintenance) on a status page

create_status_page_post_update writes

Adds a new update to an existing status page post

get_status_page_post read-only

Retrieves details of a specific status page post

list_status_page_impacts read-only

Lists available impact levels for a status page

list_status_page_post_updates read-only

Lists all updates for a specific status page post

list_status_page_severities read-only

Lists available severity levels for a status page

list_status_page_statuses read-only

Lists available statuses for a status page

list_status_pages read-only

Lists all status pages with optional filtering

create_webhook_subscription writes

Creates a new webhook subscription (returns signing secret)

delete_webhook_subscription destructive

Deletes a webhook subscription

get_webhook_subscription read-only

Retrieves a specific webhook subscription

list_webhook_subscriptions read-only

Lists webhook subscriptions with optional type filtering

update_webhook_subscription writes

Updates an existing webhook subscription

get_extension_schema read-only

Retrieves a specific extension schema

list_extension_schemas read-only

Lists available extension schemas (vendors)

Use cases

View, acknowledge, and resolve PagerDuty incidents from an IDE or chat client

Query and manage on-call schedules and create overrides

Create or update escalation policies, services, and teams

Generate incident metrics, responder load, and performance reports

Manage status page posts and updates

Use the built-in Incident Command Center, On-Call Manager, and other interactive apps inside VS Code

Supported clients

Claude Desktop

Cursor

VS Code

Listed from the project's documentation, not tested by this site.

Overview

PagerDuty's officially maintained local MCP (Model Context Protocol) server that exposes tools for MCP-enabled clients to interact directly with a PagerDuty account — viewing and managing incidents, services, on-call schedules, escalation policies, event orchestrations, status pages, and webhooks. By default the server only exposes read-only tools; write tools that can modify account data require explicitly passing --enable-write-tools. It also bundles interactive React-based apps (Incident Command Center, On-Call Manager, On-Call Compensation Report, Service Dependency Graph, Onboarding Wizard) that render directly inside VS Code. It supports stdio, streamable-http, and sse transports, and can be run via uvx, from a local checkout, or in Docker.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 62.9k · Tools 2 Compare with this →

NPM Sentinel MCP Server 79 · B

AI-powered NPM package analysis MCP server

★ 18 · Tools 19 Compare with this →

MartinLoop MCP Server 79 · B

An independent control layer that gives AI coding agents budgets, brakes, and receipts

★ 409 Compare with this →

GitHub MCP Server 76 · B

GitHub's official MCP server for managing repos, issues, PRs, and workflows via natural language

★ 33.5k · Tools 15 Compare with this →

Compare with other MCP servers

Source revision dcfbdb82e1a6 Data synced 2026-10-11 Read the FMRS scoring method