Metrics
As well as logging, the other key element of observability is metrics. Metrics allow you to see the general state of the system and observe trends and situations that are mostly caused by multiple, perhaps even many, tasks being executed at the same time.
During this chapter, we will mostly use examples of web services, like request metrics. Do not feel restricted by them; you can generate metrics in all kinds of services!
When monitoring a live system, typically metrics are the main focus, as they allow you to see at a glance whether everything appears to be working correctly. Normally with metrics, it is possible to detect if a system is struggling, for example, for a sudden increase in the number of incoming requests, but also to foresee problems by showing trends, like a small but constant increase in the number of requests. This allows you to act preemptively, without waiting until a problem is serious.
Generating a good metric system to monitor...
