Search transactions, agents…⌘K
Checking backend…
RJ

Chaos Engineering

Fault injection experiments, dependency maps, and recovery analytics

How Chaos Engineering Works

Use the Experiment Catalog to simulate live production faults. Clicking Run will inject that failure (e.g. killing the Fraud agent, slowing down DB queries) into the network. Watch the Service Dependency Map to see affected nodes turn amber (degraded) or red (failed) in real-time. After a brief duration, AegisAI's automated Self-Healing Engine will bypass or recover the service, restoring metrics to healthy states.

Total Experiments
8
Running Now
1
Completed
3
Failed
1
Experiment Catalog
Kill Fraud AgentAgent Kill
Blast: 45%Recovery: 4.2s↓12% trust
Database FailureInfrastructure
Blast: 90%
Redis Cache MissCache Failure
Blast: 30%Recovery: 1.8s↓5% trust
High Latency InjectionNetwork Delay
Blast: 60%
Prompt Injection TestAdversarial
Blast: 20%
Model Drift SimulationML Chaos
Blast: 75%Recovery: 9.6s↓28% trust
API Gateway FailureInfrastructure
Blast: 85%
Data Poisoning AttackAdversarial
Blast: 55%
Service Dependency Map
Live health state · Blast radius highlighted
healthy
degraded
failed
Before vs After Recovery
Last completed experiment: Kill Fraud Agent
Trust Score — Recovery Timeline
During Kill Fraud Agent experiment