Chaos engineering has evolved in order to increase confidence in production environments when turbulent and erratic conditions suddenly occur.
The main idea is to detect potential risk or failure points prior to those taking place and deal with them before they become cardinal issue and impact quality of service.
On a kubernetes environment, the need is even greater, having an environment and applications that can recover from failures is a great advantage and helps keep stress levels to a minimum while keeping service levels to a maximum.
Chaos kube is one of the tools available to test the resilience of a service when parts of the application are unstable.
in it’s essence , chaoskube is basically a pod that terminates other pods arbitrarily (sticking to a set of preconfigured conditions)
for a full configuration of chaoskube on openshift see my github repo