Sign In

Impact Analysis

Use the dependency graph and entity relationships to understand potential impacts during incidents or before changes.

Prerequisites

  • Catalog entities with relationships configured
  • Access to the dependency graph

1. Assess Impact During an Incident

When an incident occurs:

  1. Link the incident to affected catalog entities
  2. Open each entity’s dependency graph
  3. Look for entities connected with “depends on” relationships
  4. Check the status indicators (green, yellow, red borders)
  5. Identify services that depend on the failing component
  6. Notify teams responsible for downstream services

2. Plan for Maintenance

Before scheduled maintenance:

  1. Navigate to the entity you plan to modify
  2. Open the dependency graph
  3. Identify all services that depend on this entity (up to 2 levels shown)
  4. Coordinate with teams owning dependent services
  5. Plan your communication strategy

3. Understand Relationship Impact

Different relationship types indicate different potential impacts:

Relationship Potential Impact
Depends on Direct impact likely if this service fails
Backed by May experience degraded performance
Supports Might have reduced functionality
Parent/Child Organizational relationship, impact varies

4. Check Status Across Dependencies

From the dependency graph:

  1. Note node border colors indicating status
  2. Green borders indicate healthy services
  3. Yellow borders indicate degraded performance
  4. Red borders indicate failing services
  5. Gray borders indicate unknown status

Status from linked monitors affects entity health indicators.

Result

You can identify affected services and understand the scope of impact using the dependency graph and relationship information.

Best Practices

Document critical paths - Note which dependency chains are most important to your operations

Use entity metadata - Add descriptions and custom fields to clarify criticality and ownership

Keep relationships current - Regularly review and update dependency information as architecture changes

Create linked incidents - When multiple services are affected, create incidents for each to track resolution