Observability: It’s Not What You Think

What Is Observability?

Observability is a mindset that enables you to answer any question about your entire business through the collection and analysis of data. If you ask other folks, Observability is the dry control theory definition of “monitoring the internal state of a system by looking at its output,” or it’s the very technical definition of “metrics, traces, and logs.” While these are correct, Observability isn’t just one thing you implement, then proudly declare “now this system has Observability™.” Building Observability into your business lets you answer questions about your business.

What Kind of Questions?

Of course, the basic “what happened in our app when this error count spiked up” questions can be answered with Observability tools, but that’s barely scratching the surface of what Observability  actually is. What an Observability mindset lets you do is to figure out why the error count spiked up. If you’re intimately familiar with your app and all of its dependencies, then perhaps you can get this insight from a monitoring system, but as modern apps become increasingly more complex, the ability to maintain the state of them in your head becomes more and more challenging. Business demands, feature launches, A/B tests, refactoring into microservices… things like this all combine to create ever-increasing entropy, so knowing everything about your system without help gets more difficult by the day.