Hands-On Microservices with Kubernetes

Grudgingly accepting system failure

So, system failures will happen. Even the largest cloud providers have outages from time to time. There are different levels of system failures, ranging from temporary short failure of a non-critical subsystem, through to total outage of the entire system for a prolonged time, all the way to massive data loss. An extreme example is when malicious attackers target a company and all its backups, which can put it out of business. This is more related to security, but it's good to understand the full spectrum of system failures. Common approaches for dealing with system failures are redundancy, backups, and compartmentalization. These are solid approaches, but are expensive, and, as we mentioned earlier, will ...