Instrumenting distributed systems for operational visibility

≡ paper

If you can't look at metrics and tell whether your customers are happy, then you aren't yet monitoring the right thing. And if you can't figure out the root cause of an incident, you aren't measuring the right things either. But if you log too much noise, your observability bill can be too high. This article goes deep into how to think about instrumentation and how to turn that instrumentation into logs, metrics, and traces.

→ Read the paper
https://aws.amazon.com/builders-library/instrumenting-distributed-systems-for-operational-visibility/