Skip to main content
Tools: get_related_incidents, get_past_incidents, get_outlier_incidentCategory: IncidentsMode: Read-onlyRequires: Event Intelligence package or Digital Operations plan

​
Overview

These advanced incident analysis tools leverage PagerDuty’s Event Intelligence capabilities to help you identify patterns, find similar past incidents, and detect anomalies. This information is invaluable for faster incident resolution and understanding incident trends.
Event Intelligence RequiredThese tools require the Event Intelligence package or Digital Operations plan. If your account doesn’t have these features, these tools will return empty results or errors.
Find incidents that are currently happening and are related to the specified incident.

​
Description

Returns the 20 most recent related incidents that are impacting other responders and services. Related incidents are identified through machine learning analysis of incident metadata, service dependencies, and timing patterns.

​
Parameters

​
incident_id
string
required
The ID of the incident to find related incidents for.Example: "Q1234567890ABCDE"
​
additional_details
list[string]
Additional attributes to include in the response.Allowed values:
  • incident - Include full incident details
Example: ["incident"]

​
Usage Examples

{
  "incident_id": "Q1234567890ABCDE"
}

{
  "incident_id": "Q1234567890ABCDE",
  "additional_details": ["incident"]
}

​
Response Format

Returns a RelatedIncidentsResponse object.
​
related_incidents
list[RelatedIncident]
List of related incidents.
​
incident
PastIncidentReference
Reference to the related incident.
​
id
string
The incident ID
​
title
string
The incident title
​
created_at
datetime
When the incident was created
​
self
string
URL to access the incident
​
relationships
list[Relationship]
List of relationship types to the parent incident.
​
type
string
Relationship type (e.g., machine_learning_inferred, service_dependency)
​
metadata
dict
Additional metadata about the relationship (structure varies by type)

​
Use Cases

When to use related incidents:
  • During major incidents - Identify if other systems are affected
  • Investigating cascading failures - Find upstream/downstream impacts
  • Coordinating response - See what other teams are working on
  • Understanding scope - Determine if issues are isolated or widespread
  • Post-incident review - Map out full incident timeline across services

​
get_past_incidents

Find historical incidents similar to the current incident.

​
Description

Returns incidents from the past 6 months that have similar metadata and were generated on the same service. By default, returns 50 past incidents with similarity scores. This helps you learn from previous resolutions and identify recurring patterns.

​
Parameters

​
incident_id
string
required
The ID of the incident to find past incidents for.Example: "Q1234567890ABCDE"
​
limit
integer
default:"50"
Number of results to return. Must be between 1 and 999.Example: 100
​
total
boolean
default:"true"
Include the total count of past incidents in the response.Example: true

​
Usage Examples

​
Find past similar incidents

{
  "incident_id": "Q1234567890ABCDE"
}

​
Get more past incidents

{
  "incident_id": "Q1234567890ABCDE",
  "limit": 100,
  "total": true
}

​
Response Format

Returns a PastIncidentsResponse object.
​
past_incidents
list[PastIncident]
List of past incidents sorted by similarity.
​
incident
PastIncidentReference
Reference to the past incident.
​
id
string
The incident ID
​
title
string
The incident title
​
created_at
datetime
When the incident was created
​
self
string
URL to access the incident
​
score
float
Similarity score between 0 and 1, where higher scores indicate greater similarity.
​
total
integer
Total number of past incidents found (if total parameter was true).
​
limit
integer
The limit that was applied to the results.

​
Use Cases

When to use past incidents:
  • New responders on-call - Quickly find how similar issues were resolved before
  • Investigating unfamiliar issues - Learn from past resolutions
  • Identifying recurring problems - Spot patterns that need permanent fixes
  • Estimating resolution time - See how long similar incidents took
  • Finding relevant runbooks - Check notes from past similar incidents
  • Post-incident analysis - Compare with historical data

​
get_outlier_incident

Identify if an incident is an outlier for its service.

​
Description

Analyzes an incident to determine if it deviates from expected patterns for the same service. Outlier incidents are rare or unusual compared to the normal incident patterns, which may indicate a novel issue requiring different handling.

​
Parameters

​
incident_id
string
required
The ID of the incident to analyze.Example: "Q1234567890ABCDE"
​
since
datetime
The start of the date range to analyze. Maximum range is 6 months.Example: "2024-01-01T00:00:00Z"

​
Usage Examples

​
Check if incident is an outlier

{
  "incident_id": "Q1234567890ABCDE"
}

​
Check outlier status with date range

{
  "incident_id": "Q1234567890ABCDE",
  "since": "2024-01-01T00:00:00Z"
}

​
Response Format

Returns an OutlierIncidentResponse object.
​
outlier_incident
OutlierIncident
Outlier incident analysis results.
​
incident
OutlierIncidentReference
The incident that was analyzed.
​
id
string
The incident ID
​
title
string
The incident title
​
created_at
datetime
When the incident was created
​
occurrence
Occurrence
Occurrence pattern information.
​
count
integer
Number of times this pattern has occurred
​
frequency
float
Frequency of occurrence
​
category
string
Occurrence category (e.g., “rare”, “common”)
​
since
datetime
Start of analysis time range
​
until
datetime
End of analysis time range
​
incident_template
IncidentTemplate
Pattern template information.
​
id
string
Template ID
​
cluster_id
string
Cluster ID for grouping similar incidents
​
mined_text
string
The mined text pattern that characterizes this incident type

​
Use Cases

When to use outlier incidents:
  • Triaging incidents - Prioritize unusual incidents that may indicate new problems
  • Anomaly detection - Identify incidents that don’t follow normal patterns
  • Investigating unknowns - Focus attention on truly novel issues
  • Resource allocation - Assign experienced responders to outlier incidents
  • Pattern analysis - Track when new incident patterns emerge
  • Capacity planning - Understand what percentage of incidents are unique

​
Common Errors

Event Intelligence not enabledIf your account doesn’t have Event Intelligence:
402 Payment Required: Event Intelligence features require an Event Intelligence package or Digital Operations plan
Solution: Contact PagerDuty to upgrade your plan.
Incident not found
404 Not Found: Incident not found
Solution: Verify the incident ID is correct.
Date range too largeFor get_outlier_incident:
400 Bad Request: Date range cannot exceed 6 months
Solution: Use a since date within the last 6 months.

​
Combining Tools for Incident Analysis

These tools work well together for comprehensive incident analysis: 
# 1. Get the incident details
get_incident({"incident_id": "Q1234567890ABCDE"})

# 2. Check if it's an outlier
get_outlier_incident({"incident_id": "Q1234567890ABCDE"})

# 3. Find related ongoing incidents
get_related_incidents({"incident_id": "Q1234567890ABCDE"})

# 4. Find similar past incidents for resolution hints
get_past_incidents({"incident_id": "Q1234567890ABCDE", "limit": 10})

get_incident

Get the main incident details

list_incident_notes

Check notes from past similar incidents

manage_incidents

Take action on related incidents

list_incidents

Find other recent incidents

​
Source Reference

Functions:
  • get_related_incidents in pagerduty_mcp/tools/incidents.py:280
  • get_past_incidents in pagerduty_mcp/tools/incidents.py:259
  • get_outlier_incident in pagerduty_mcp/tools/incidents.py:239
Models:
  • RelatedIncidentsQuery, RelatedIncidentsResponse in pagerduty_mcp/models/incidents.py
  • PastIncidentsQuery, PastIncidentsResponse in pagerduty_mcp/models/incidents.py
  • OutlierIncidentQuery, OutlierIncidentResponse in pagerduty_mcp/models/incidents.py

Build docs developers (and LLMs) love

Get started for freeTalk to us