Traces to metrics: Perform ad hoc queries in Grafana Tempo

In the realm of observability, Grafana Tempo's updated features, including "Aggregate by," enable more effective trace analysis by allowing users to compute RED (Rate, Error, Duration) metrics at query time rather than storing massive amounts of data in a time series database. This approach helps avoid the pitfalls of cardinality explosion, which can be costly and inefficient, by facilitating on-the-fly metrics computation tailored to specific investigative needs. The "Aggregate by" feature enables users to break down traces by various attributes, providing a granular view of system performance and aiding in pinpointing issues like elevated error rates or latency confined to particular service regions or endpoints. These advancements are part of a broader strategy to integrate ad hoc metrics computation directly into TraceQL, offering greater flexibility and insight without the overhead of persistent metrics storage. Early adopters like Sarwa have praised the feature for its ability to swiftly identify problematic endpoints, demonstrating its practical utility in real-world scenarios. Grafana Labs is committed to expanding these capabilities, aiming for more comprehensive trace metrics analysis in future updates.