Tracing is a specialized form of logging. It provides lower-level information than logs. This may include all the function calls, their parameters, their size, and execution time. They also contain the unique ID of the transaction being processed. These details make it possible to reassemble them and see the life cycle of a given transaction as it passes through your system.
Performance information present in tracing helps you with uncovering bottlenecks and sub-optimal components in the system.
While logs are often read by operators and developers, they tend to be human-readable. There are no such requirements for tracing. To view the traces, you will use a dedicated visualization program. This means that even though traces are more detailed, they may also take up less space than logs.
The following diagram is an overview of a single trace:
Two services communicate over a network. In Service A, we have one...
