Distributed application tracing

Distributed and networked systems are inevitable in any large-scale infrastructure. That's because you can put only a limited amount of hardware in a single server rack. If you have to serve a large number of users concurrently, you will have to scale out eventually. Also, having all of your software running on a single host will be risky from a reliability and availability standpoint. If one machine fails, the whole system goes down. And it isn't that uncommon to see even whole datacenters being brought down due to natural disasters or other unpredictable events. This means that highly available systems are often forced to be spread over multiple datacenters located in different regions or even managed by different providers, just to ensure enough redundancy.

Another reason for distributed systems is splitting infrastructures into independent domain services in order to split large codebases and enable multiple teams to work efficiently...