Distributed tracing in Linkerd is automated and does not require any special configuration. The data that's collected from the Linkerd proxy is aggregated to show health, latencies, and request volumes at the service level and path/route level.

The dashboard also shows a live dependency graph and the topology of the services. There are four golden metrics for monitoring distributed systems; that is, Latency, Traffic (RPS), Errors, and Saturation, which an SRE team will be very interested in getting a report on. The Linkerd dashboard shows these golden metrics through its dashboard. The following screenshot shows five metrics for a web app:

Distributed tracing requires applications to preserve request headers if they are set by the proxy. Future releases of Linkerd may use distributed tracing through span headers. Istio supports application tracing through span headers already, but it lacks in showing the golden metrics, as shown by Linkerd. 

The P50, P95, and P99 latency metrics can help application developers identify application bottlenecks. This aggregated latency information is a very useful feature of Linkerd.