Operating OpenShift

Chapter 7. Advanced Monitoring and Observability Strategies

Rick Rackow

In Chapter 6 you got to know OpenShift’s monitoring stack and the basics of monitoring, alerting, and observability. This chapter looks further into this topic and covers more advanced strategies to monitor your services.

Service Oriented Monitoring

Just as infrastructure and services have changed over time, so has the approach to monitoring, and not just since the publishing of Site Reliability Engineering by Betsy Beyer and colleagues have engineers sought better ways of monitoring their stack. With the shift to containerized workloads in particular, there are significantly more moving pieces of infrastructure, and it is not feasible to treat each of them with minute care. Think of it as cattle vs. pet. Each of your pets has a name, and you know exactly how it is doing and what it possibly needs. In cattle, each animal is part of the herd, and they’re more than merely a number to their owner. However, what matters is not so much the feeling of each individual animal but much rather that the results add up, like the amount of milk received in a day from the cattle. In practice, that means instead of checking the health of underlying nodes, or even worse, individual containers, you shift your focus to the health of your services. Though that might feel strange at the beginning, the benefits become more obvious the more your infrastructure and service landscape grows.

If you consider a basic OpenShift installation ...