Skip to content

Risk Scoring

FaultRay calculates a resilience score from 0 to 100 based on multiple weighted factors. This page explains the scoring algorithm and how to interpret results.

Score Breakdown

The overall resilience score is a weighted combination of four sub-scores:

Factor Weight Description
SPOF Score 30% Penalty for single points of failure
Cascade Score 25% Risk of cascading failures
Redundancy Score 25% Coverage of redundant paths
Geographic Score 20% Distribution across regions/AZs

SPOF Score (30%)

Single Points of Failure receive the highest weight because they represent the most critical risk.

spof_score = 100 - (spof_count * spof_severity_weight)

Each SPOF is weighted by its position in the dependency graph:

  • Root-level SPOF (e.g., single database): -30 points
  • Mid-tier SPOF (e.g., single app server behind LB): -15 points
  • Edge SPOF (e.g., single CDN origin): -10 points

Cascade Score (25%)

Evaluates how far failures propagate through the dependency graph.

cascade_score = 100 - (max_cascade_depth / total_depth * 100)

A cascade depth of 1 (isolated failure) is ideal. A cascade that reaches all layers indicates severe architectural risk.

Redundancy Score (25%)

Measures the percentage of components that have redundant alternatives.

redundancy_score = (redundant_components / total_components) * 100

Components are considered redundant if they have:

  • Active-active or active-passive failover peers
  • Auto-scaling groups with min instances > 1
  • Cross-region replicas

Geographic Score (20%)

Evaluates distribution across failure domains (regions and availability zones).

geo_score = (unique_azs / recommended_azs) * 100

Best practice is a minimum of 3 AZs for production workloads.

Score Interpretation

Score Grade Action Required
90-100 A Production-ready. Monitor and maintain.
80-89 B Good. Address minor improvements.
60-79 C Fair. Plan remediation for identified risks.
40-59 D Poor. Immediate attention required.
0-39 F Critical. Do not deploy without remediation.

Customizing Weights

Override default weights in your configuration:

scoring:
  weights:
    spof: 0.35
    cascade: 0.25
    redundancy: 0.25
    geographic: 0.15

faultray simulate -m model.json --scoring-config scoring.yaml