Hands-On Microservices with Kubernetes

Taking human factors into account

Now, we're strictly in the domain of people responding to an actual incident. Some critical systems may have 24/7 live monitoring by people watching the system state diligently and ready to act. Most companies will have alerts based on various triggers. Note that, even if you have live 24/7 monitoring for complex systems, you still need to surface alerts to the people monitoring the system because, for such systems, there is typically a huge amount of data and information that describe the current state. Let's look at several aspects of a reasonable alerting plan that work well for people.