Hi Reka, I'm currently implementing actions what needs to be done if a VM is down. VM down in the sense if the agent does not send any response for a given time period.
CEP check if a agent does not send any statistic for a 1 minutes. Then it emits a MemberFault event to Auto Scaler. Then auto scaler tell CC to kill that member and spawn another instance in the same partition the killed VM was. More fault handling scenarios will be implemented soon such as network of agent down,VM up but application crashed...etc Touched, not typed. Erroneous words are a feature, not a typo.
