- 9. Sep
Chaoskube is an open source Chaos Testing tool. chaoskube periodically kills random pods in your Kubernetes cluster.Chaos Engineering is the discipline of proving the reliability of any system by causing “chaos”. The work ‘Chaos’ means the state of confusion or failure caused due to unexpected reason.
Failures can be caused due to:
- Power outages.
- Software bugs.
- Human Error.
Since failure is unavoidable.
Why not deliberately introduce failure to ensure the system can deal with the failure?
Chaoskube is one such tool, which can be used to introduce pod failures on Kubernetes Cluster.
Overview of Chaoskube:
- Chaoskube is an open source Chaos Testing tool.
- Written in GO language.
- Can induce pod/controller failures on K8s Cluster.
- Can kill pods by specifying the labels, namespaces.
- Simple and easy to run.
Induce controller failure using Chaoskube:
- Induce failure on pod with label ‘openebs/controller=jiva-controller’ for duration of 60 seconds with interval of 20 seconds, which means it will induce controller pod
failure for every 20 seconds for 3 times.
- Observe that percona application pod with liveness probe is still running after inducing openebs controller pod failure using chaoskube. Hence, the system is reliable
after causing ‘Chaos’.