Chapter 7. Visualization

If you want to summarize the intent of systems monitoring in two words, it would be difficult to do better than increase visibility. There are too many systems doing too many things too quickly for humans to maintain any sense of what’s going on at any given moment. The best we can hope for is to tell when something breaks, but a good monitoring system isn’t bound by our organic limitations, so it shouldn’t stop there. Good monitoring systems act like transducers in electronics, converting the incomprehensibly large number of interactions between systems and networks into an environmental compendium fit for human consumption. Good monitoring systems provide organic interfaces to the network, which allow us to see more, and the extent to which they accomplish this determines their usefulness as tools.

Humans are visual animals, so the best way to understand complex information is to draw us a picture. The importance of data visualization to the success of your implementation and the well-being of your environment cannot be understated. Good visualization solves the visibility problem by effectively communicating the status of the environment, enabling pattern recognition in historical and real-time data, and making propeller-head metrics into indicators laymen can utilize. Good visualization can aid any number of critical undertakings such as capacity planning, forensics, and root cause analysis. Visualization catches problems you didn’t tell the monitoring system to look for. On the other hand, absent or poor data visualization severely hinders the usefulness of the system and can render your monitoring system irrelevant altogether.