ML-Driven Chaos Experiments
Quick Reference
What & Why
Definition
>= 60% of chaos experiments use ML to select targets, predict blast radius, auto-tune intensity for maximum learning.
Business Value
Generates 70% of test cases automatically with 90% assertion quality and discovers 3x more edge cases through AI-powered test generation Achieving >= 60% chaos experiments ML-driven is a key milestone toward this goal.
Context
This capability is part of the Optimization milestone's focus on ai enablement, predictive ops, self-healing. Essential for teams targeting CFR, LT improvements.
Success Criteria
>= 60% chaos experiments ML-driven
Measurement
ML chaos experiment coverage
Evidence
- Chaos ML model
- Experiment results
- Blast radius predictions vs actual
In Practice
Real-World Implementation
ML selects chaos target (high-impact service), predicts blast radius, sets failure intensity, monitors impact, auto-halts if exceeds prediction. Maximizes learning, minimizes risk.
Concrete Example
Implementation Guide
Implementation Steps
Follow the measurement approach: ML chaos experiment coverage
For detailed step-by-step guidance, refer to the AI-Generated Testing & Intelligent Quality Implementation Kit.
Resources
Implementation Kit
AI-Generated Testing & Intelligent Quality KitTemplates
Browse all templatesRelated Resources
View learning pathsRelated Capabilities
Prerequisites
Implement these first
Complementary
Often adopted together, from the AI-Generated Testing & Intelligent Quality epic
Troubleshooting & FAQs
Common Issues
Issue: Target metric not improving
Solution: Verify measurement is accurate, check if prerequisites are fully implemented, review evidence artifacts for completeness
Issue: Team resistance to adoption
Solution: Start with pilot team, demonstrate value with metrics, provide training and support during transition
Issue: Inconsistent implementation across teams
Solution: Create shared templates and guidelines, establish regular sync meetings, use automation to enforce standards
Frequently Asked Questions
Can we implement this before completing prerequisites?
While possible, it's not recommended. Prerequisites ensure foundational practices are in place, making this capability more effective and easier to adopt.
How long does implementation typically take?
Most capabilities can be implemented within 185 days when tackled as part of the Optimization milestone. Individual timelines vary based on team size and existing practices.