pagerduty · MCP Server

PagerDuty + Improvado MCP — Incident Intelligence in Plain English

Improvado's MCP server pulls PagerDuty incident data into your AI agent. Query open incidents, escalation history, MTTR trends, and on-call coverage — without logging into PagerDuty every time. Works with Claude, Cursor, and any MCP-compatible tool.

46K+ metrics ·Read & Write access ·500+ platforms ·<60s setup

Read

Read: Get Incident Answers Instantly

Stop clicking through incident timelines and service directories. Ask your AI agent for MTTR by team, recurring alerts by service, or on-call coverage gaps. The MCP server handles the PagerDuty API so you don't have to.

Example prompts

"Which services triggered the most P1 incidents in the last 30 days? Break down by team."

25 min → 30 sec

"Show me average MTTR for the payments service over the last 90 days."

20 min → 20 sec

"Who's on call this weekend across all production services?"

10 min → 10 sec

Works with Claude ChatGPT Cursor +5

Write

Write: Manage Incidents Without Switching Tabs

Acknowledge incidents, add notes, reassign responders, and update escalation policies directly through your AI agent. Critical incident actions in one prompt — no more racing through the PagerDuty UI under pressure.

Example prompts

"Acknowledge all open P2 incidents for the API gateway service and add note: investigating load spike."

5 min → 20 sec

"Escalate incident #INC-4821 to the senior on-call engineer for the payments team."

3 min → 15 sec

"Create a maintenance window for the database cluster every Sunday 2–4am."

15 min → 1 min

Every action logged · Fully reversible · SOC 2 certified

Monitor

Monitor: Catch Patterns Before They Become Outages

Set up AI-powered watches on incident frequency, MTTR drift, and alert noise. Get notified about degrading reliability trends before they show up in post-mortems.

Example prompts

"Alert me if any service's incident rate doubles week-over-week."

Manual → auto

"Every Monday: send a summary of MTTR by team and top recurring alerts."

2 hrs → auto

"Flag services with more than 10 incidents in 7 days that have no post-mortem linked."

Manual → auto

Alerts sent to Slack, email, or your AI agent

Full cycle

The Closed Loop: Read → Decide → Write → Monitor

Your AI agent doesn't just surface data — it acts. Adjust pricing, update product descriptions, manage inventory, apply discounts — all through natural language. The MCP server translates intent into API operations.

Every phase runs through the same MCP connection. One protocol, all platforms, full governance. No switching between tools.

Ideate

Launch

Measure

Analyze

Report

Iterate

One conversation. All six phases. Every platform.

The daily grind

Common problems. Direct answers.

Challenge 1

Post-Mortem Data Lives in Scattered Places

The problem

After a major incident, the timeline is split between PagerDuty alerts, Slack threads, and a Confluence doc someone partially filled in. Reconstructing the exact sequence of events — who was paged, when they acknowledged, what actions were taken — takes hours of manual archaeology.

How MCP solves it

Improvado's MCP server pulls the full incident timeline from PagerDuty: notification times, acknowledgment gaps, escalation steps, and resolution notes. Your AI agent assembles the chronology in seconds, ready to paste into a post-mortem template.

Try asking

Give me the full timeline for incident #INC-5042: who was paged, when they acknowledged, all escalations, and resolution time.

Answer in seconds

All data sources, one query

Challenge 2

Alert Fatigue Hides Real Signal

The problem

Services fire hundreds of low-priority alerts weekly. Engineers stop reading them. Then a real P1 gets buried in noise and goes unacknowledged for 40 minutes. You know you need to tune your alert policies, but identifying the noisy ones means exporting CSVs and doing the analysis yourself.

How MCP solves it

Ask your AI agent to analyze alert frequency by service and policy. It identifies which services generate high-volume low-priority noise versus which ones escalate to real incidents. Tune policies based on actual data, not guesswork.

Try asking

Which alert policies fired more than 50 times last week but had zero P1 escalations? Rank by noise volume.

Full detail preserved

No data loss on export

Challenge 3

On-Call Coverage Gaps Go Unnoticed Until It's Too Late

The problem

Someone forgot to update the on-call schedule during a holiday week. A critical service has no primary responder assigned. Nobody notices until 2am when an incident fires and nobody gets paged. Auditing schedules manually across 20 services is tedious enough that it rarely happens.

How MCP solves it

Your AI agent audits PagerDuty schedules proactively. It surfaces gaps, services with a single point of failure on-call, and rotation overloads where one engineer is covering too many services simultaneously.

Try asking

Show me any on-call coverage gaps in the next 14 days across all production services.

Unified data model

Compare anything side by side

👥 Teams

One Framework. Five Roles. Zero Setup.

Same MCP connection, different workflows for every team member. Each role asks in natural language — the MCP server handles the complexity (rate limits, auth, schema normalization, governance) behind the scenes.

Agency CEO

Portfolio health. Client risk. Revenue signals.

Media Strategist

70% strategy, not 70% ops. Auto campaign QA.

Marketing Analyst

Zero wrangling. Cross-platform. AI narratives.

Account Manager

QBR decks auto-generated. Call prep in 30s.

Creative Director

Performance-to-brief. Predict winners before spend.

FAQ

Common questions

What PagerDuty data can I query through the MCP server?

Incidents (status, severity, timeline, responders, notes), services, escalation policies, on-call schedules, and alert frequency metrics. You can query across all services at once or drill into a specific team's incident history.

Can I write back to PagerDuty — not just read data?

Yes. Acknowledge and resolve incidents, add notes, reassign responders, create maintenance windows, and update escalation policies — all through natural language. Write permissions are scoped to your PagerDuty API token.

Does this work with PagerDuty's team and service hierarchy?

Yes. The MCP server understands PagerDuty's structure — services, teams, escalation policies, and schedules. You can query by team, by service, or across the whole organization and the results respect your existing hierarchy — all through Improvado's hosted MCP server.

How is this different from PagerDuty's built-in analytics?

PagerDuty's analytics require you to navigate their UI and work within their reporting templates. With Improvado MCP, you ask ad-hoc questions in plain English, combine PagerDuty data with data from other tools in the same query, and get instant answers without building reports.

Can the PagerDuty MCP integration be used for post-incident analysis and retrospectives?

Yes, the PagerDuty MCP integration is well-suited for post-incident analysis. An AI agent can retrieve full incident timelines, acknowledgment and resolution timestamps, escalation paths, and associated alert details to help reconstruct the sequence of events during an outage. You can ask questions like 'what was the mean time to resolution for P1 incidents in the last quarter' or 'which services had the most escalations last month' to surface patterns without manually exporting and analyzing PagerDuty reports — all through Improvado's hosted MCP server.

Does the PagerDuty MCP integration support querying on-call schedules and team coverage?

Yes, the PagerDuty MCP integration can query on-call schedule data, escalation policy configurations, and team assignments through PagerDuty's Schedules and Oncalls API. This allows an AI agent to answer questions like 'who is on call for the payments service this weekend' or 'identify gaps in on-call coverage for next week' without navigating through multiple PagerDuty screens. This is useful for engineering managers and SRE leads planning rotations or auditing coverage before major deployments — all through Improvado's hosted MCP server.

Stop Reporting. Start Executing.

Connect your data to an AI agent in under 60 seconds. The closed loop starts with one conversation.

SOC 2 Type II GDPR 500+ Platforms