Add a scenario to randomly kill pods based on a regexp

[mffiedler]

Add a scenario to kill a random pod selected from pods matching a regexp. We have a simple, but effective shell script in OpenShift testing that randomly takes out infra pods. Combined with Cerberus (https://github.com/openshift-scale/cerberus) we can detect non-transient issues caused by such a scenario.

Example (in shell scripting):

#!/bin/bash
NUM=$1
SLEEP=$2
while true; do 
   echo "============"
   oc get pods --all-namespaces | grep openshift | grep Running| awk {'print $1" "$2'} | shuf -n $NUM | while read i; do oc delete pod -n $i --wait=false; done; sleep $SLEEP; 
done

[chaitanyaenr]

Fixed by #16. Thanks to @paigerube14.