Netflix created Chaos Monkey to randomly terminate instances. The Simian Army extends this concept.
Types of fault injection:
- Instance termination: Kill random servers
- Network latency: Add delay to calls
- Network partition: Block communication between services
- Disk fill: Exhaust storage
- CPU stress: Simulate resource contention
- DNS failure: Break name resolution
Tools:
- Chaos Monkey (Netflix)
- Gremlin (commercial)
- LitmusChaos (Kubernetes)
- AWS Fault Injection Simulator