Historically, google has chosen to leave these problems to the service layer. Their internal systems detect these events but don’t readily remediate them. I think this is great policy, it’s the reason why Google’s stack has excellent, sophisticated client-side load direction and error avoidance.