..
Observability Concepts
- Traces: A trace provides a view of how a request travels through a distributed system. For example, it might show the journey of a request as it moves from the front end to the back end, interacts with a database, and returns to the front end. This helps identify bottlenecks or failure points across different components. 1
- Spans: A span represents a single operation within a trace, with information on when the operation started and ended. Spans are the building blocks of a trace, as each span details an individual step within the request’s journey, enabling precise insights into which part of the process might be lagging.
- Metrics: Metrics can represent anything quantifiable within the system, such as the request processing time or the number of requests over a period. Metrics are crucial for tracking performance trends over time and can help detect unusual behavior, like a sudden spike in response times or request volume.