Gzchaos [2026]

gzchaos inject network loss \ --target pod/frontend-7d8f9b6c4d-abc12 \ --namespace prod \ --duration 30s \ --loss-percent 100 \ --dry-run # preview without execution Confirm injection:

gzchaos revert --experiment-id <id> | Fault Type | Example Use | Key Parameters | |------------|--------------|----------------| | Latency | Simulate slow downstream DB | --latency 500ms --jitter 50ms | | Error injection | Test HTTP 500 / gRPC UNAVAILABLE | --error-code 500 --rate 0.3 | | Pod kill | Test controller rescheduling | --grace-period 0 --restart-policy never | | CPU stress | Trigger HPA scale‑up | --cores 2 --duration 60s | | Disk pressure | Simulate full tmpfs | --fill-mb 1024 --path /scratch | | DNS failure | Test fallback IP / cache | --domain api.example.com --nxdomain | 6. Safety & Guardrails Enable safeguards via config or flags: gzchaos

blast_radius: max_targets: 3 allowed_namespaces: ["staging", "chaos-engineering"] excluded_labels: ["tier=control-plane"] auto_stop: on_latency_spike: ">2s" on_error_rate: ">25%" on_rollback_signal: true approval: require_manual_ok: true # for prod experiments gzchaos schedule apply --file weekly-chaos.yaml Example weekly-chaos.yaml : 2s" on_error_rate: "&gt

gzchaos experiment run --safety-config safety.yaml safety.yaml example: gzchaos

gzchaos status --experiment-id <id> Rollback immediately: