📊 SRE & Platform Engineering — Chaos Engineering Toolkit Demo

• Failure Injection Engine — Decorator-based latency/error/resource injection with safety controls
• Steady-State Hypothesis Checker — Define and evaluate system health before, during, and after experiments
• Blast Radius Control — Progressive expansion with consistent hashing, automatic abort, and kill switch
• Experiment Catalog — Pre-built YAML experiment definitions for common failure scenarios
• Game Day Templates — Structured planning documents for team chaos exercises
• Maturity Model Guide — Five-level progression framework for building a chaos program

# Review available experiments
ls experiments/

# Copy and customize for your environment
cp experiments/latency_injection.yaml my_experiment.yaml
# Edit target_service, namespace, thresholds, etc.


python -m unittest tests.test_chaos -v

from src.failure_injector import FailureInjector, InjectionConfig, FailureType

injector = FailureInjector()
config = InjectionConfig(
    failure_type=FailureType.LATENCY,
    probability=0.3,         # Affect 30% of calls
    latency_ms=500,          # Add 500ms delay
    duration_seconds=300,    # Run for 5 minutes
    blast_radius_pct=30.0,   # Safety: never exceed 30%
)
injector.activate(config)

@injector.inject_latency(config)
def call_user_service(user_id: str) -> dict:
    """This function will randomly be 500ms slower."""
    return http_client.get(f"/users/{user_id}")