Grafana has introduced enhancements to its Agent, initially launched as a subset of Prometheus for hosted metrics, by integrating Grafana Tempo for distributed tracing. The Agent, which is known for reducing memory usage by 40%, now includes features such as a clustering mechanism, additional Prometheus exporters, and support for Loki, alongside the newly added tracing capabilities. Users can configure the Agent to collect and ship traces to Tempo by adding a tempo block to their configuration file, enabling it to receive spans from various systems like Jaeger, Kafka, OpenCensus, OTLP, and Zipkin. This setup allows operators to manipulate tags on incoming spans and utilize Prometheus Service Discovery mechanisms to attach metadata, thereby simplifying the transition between different observability data types. Remote write configurations control where collected spans are sent, with options for batching, queueing, and retrying to optimize network usage and avoid memory issues. The integration aims to streamline observability across metrics, logs, and traces, with Grafana Cloud offering free and paid plans for users to leverage these capabilities.