Tracing with the Grafana Cloud Agent and Grafana Tempo
Blog post from Grafana Labs
Grafana has introduced enhancements to its Grafana Agent, a subset of Prometheus, to support distributed tracing through the integration of Grafana Tempo. This development facilitates the collection and transmission of tracing data from multiple systems such as Jaeger, Kafka, OpenCensus, OTLP, and Zipkin. The integration allows operators to manipulate metadata on incoming data spans, providing additional context and enabling dynamic tagging with metadata such as environment details. Additionally, the agent supports Prometheus-style scrape_configs for automatic tagging based on discovered targets, especially beneficial for Kubernetes users. The remote_write configuration allows spans to be sent to Tempo or other compatible systems, with options for batching and retry handling to optimize network usage. This setup is part of a broader effort to unify observability data across metrics, logs, and traces, enhancing the ease of data transition and reducing cognitive overhead. Users can explore the production-ready Kubernetes manifest for practical implementation or utilize the free and paid Grafana Cloud plans to test and scale their tracing capabilities.