[EPIC] Implement a chaos summary reporter from long running chaos testing

[davidkel]

The report creator must be able to recognise and ignore or report as expected problems which we know would happen under certain conditions Actions are going to cause problems, eg

• kill leader orderer
• take down 3 or more orderers
• take down gateway peer

These will generate errors, but what we are looking for is client recovery without having the re-start the client or the gateway peer

[davidkel]

Further consideration will come when bringing peers down. Some scenarios are not expected to cause problems but others are and also it will depend on the endorsement policy being used, so how would a report generator know if a particular scenario will cause a problem or not