CP Massive Creative

The USE pattern, developed by #Netflix performance engineer Brendon Gregg, is a bottom-up approach that is intended to help analyze performance issues and find bottlenecks. USE stands for Utilization, Saturation, and #Errors. #kubernetes

The Requests-Errors-Duration (RED) pattern is a classic observability tool that goes back to the earliest days of online services. Google’s Site Reliability Engineering book talks about the Four Golden Signals, which are essentially requests, #errors, #duration, and #saturation.

Closed-box monitoring checks observe the external behavior of a system to detect predictable failures. #observability #kubernetes

#Observability is the union of traditional monitoring, logging, metrics, and #tracing, and all the other ways you can understand your system. #kubernetes

Distributed systems expose the limitations of traditional #monitoring because they’re not in either up or down states: they exist in a constant state of partially degraded service. In other words, nothing is ever completely right aboard a ship. #kubernetes

Suppose you have a chain of 10 #microservices, each of which depends on the next for some critical part of its work. The last service in the chain fails. The next- to-last service will detect this and start failing its readiness probe. #Kubernetes

#Cloudnative applications fail in complex, and hard- to-detect ways. Applications have to be designed to be resilient and degrade gracefully in the face of unexpected failures, the harder it is to detect these failures by closed-box monitoring. #kubernetes

Perhaps the most useful way to distinguish these terms is to say that #monitoring tells you whether the system is working, while #observability prompts you to ask why it’s not working, along with considering how well it is performing. #kubernetes

Because the term #monitoring means different things to different people - from plain old closed-box checks to a combination of metrics, logging, and tracing - it’s becoming common to use #observability as a catch-all term that covers all these techniques. #kubernetes

Distributed systems like #cloudnative applications are never up; they exist in a constraint state of partially degraded service. Gray failures are, by definition, hard to detect, especially from a single point of view or with a single observation. #kubernetes #observability