When a single user request travels through many different microservices, what specific observability tool helps an expert see the full path of the request and how long each service took?

The specific observability tool that helps an expert see the full path of a single user request across many different microservices and how long each service took is Distributed Tracing. Distributed Tracing is a technique used to monitor requests as they flow through a distributed system, providing visibility into the end-to-end transaction. When a user request first enters the system, a unique identifier, known as a Trace ID, is generated for that specific request. This Trace ID is then propagated, meaning it is passed along, with the request as it travels from one microservice to the next, ensuring all related operations share the same identifier. Each distinct unit of work or operation performed by a microservice in response to this request is recorded as a Span. A Span represents a single, logical operation within a service, such as handling an incoming API call, performing a database query, or making an outgoing request to another microservice. Each Span includes key information, most importantly its unique identifier, the Trace ID it belongs to, a start timestamp, and an end timestamp. The difference between the start and end timestamps precisely indicates the duration of that specific operation. Spans are organized in a hierarchical relationship to reflect the sequence of operations. When one service calls another, the Span representing the call in the first service acts as the Parent Span for the Span representing the operation in the second service. This parent-child relationship is established by including the Parent Span's ID within the child Span's data, along with the main Trace ID. The complete set of all interconnected Spans, linked by their shared Trace ID and parent-child relationships, forms an entire Trace. This Trace visually maps out the entire journey of the single user request, showing every service it interacted with and the sequence of those interactions. By analyzing the collected Trace data, particularly the start and end timestamps of each individual Span, an expert can precisely determine how long each specific service or operation took to complete its part of the overall request, identifying bottlenecks and understanding the complete execution flow. For example, a Trace for a checkout request might include a Span for the "frontend service processing form," followed by a child Span for "order service creating order," which in turn has a child Span for "inventory service deducting stock," with each Span's duration clearly recorded.