The current approach is only useful for services delivered over http. We are sending an individual email per fault detected - we can improve this.