Deploying a tracing system is not just a matter of instrumenting your code or running a tracing backend. In this chapter, we discussed five problem areas that must be addressed in order to "have a good time" when collecting traces, and being able to analyze your system behavior and performance. The areas are analyzing, recording, federating, describing, and correlating transactions.
Most existing tracing systems cover all those areas, but in their own ways, which are incompatible with other tracing systems. This limits interoperability, which is especially problematic when using managed cloud services. There are four standardization projects that exist in the industry and they are trying to address different problem areas. We reviewed the scope of each project and discussed why having a narrow scope is important for them to be successful.
In the next chapter, we will discuss how other technologies, specifically service mesh proxies, can be used to standardize methods of extracting...