Mastering Distributed Tracing

Summary

In this chapter, we discussed how the metadata propagation mechanism can be separated from the tracing instrumentation that depends on it, for example, using the approach from the Tracing Plane, and why this is not always done in practice. We reviewed a number of cross-cutting techniques and tools for solving the problems of monitoring, debugging, and testing distributed systems that are not dependent on tracing directly, but depend on distributed context propagation. Having tracing instrumentation in the distributed systems makes these additional tools much easier to implement.

We briefly touched upon using traffic labeling to affect application metrics and alerts. In the next chapter, we will cover that in more detail, as well as other ...