Chaos Engineering
Chaos engineering proactively injects failures (killing nodes/adding latency/cutting network) to verify system resilience and uncover hidden weaknesses.
Purpose
Proactively discover system weaknesses and harden before real failures.
Mechanism
Define a steady-state hypothesis (e.g., availability>99.9%). Inject failures (kill Pods/network latency/CPU stress). Verify the steady state holds. Implemented by Chaos Monkey/Chaos Mesh/Litmus. Game days run whole-team drills.
Use cases
- Resilience validation
- Failure drills
- System hardening