Tempo 3.0 release: a new architecture for scale and lower TCO, TraceQL metrics GA, and more

Tempo 3.0 introduces a significant architectural overhaul aimed at enhancing scalability and reducing the total cost of ownership for distributed tracing, alongside the general availability of TraceQL metrics for deriving insights directly from trace data. The new Kafka-compatible architecture for microservices deployments allows Tempo to efficiently handle trace ingestion, storage, and querying by decoupling the read and write paths, which improves performance and reliability. Key improvements include enhanced query performance, trace redaction capabilities, and the ability to run TraceQL metrics without relying on the metrics-generator component. Additionally, the release introduces a per-label cardinality limiter and a DRAIN-based span name sanitizer to manage cardinality and ensure system stability, as well as flexible filtering options and sampling compensation to improve metric accuracy. Users are advised to refer to the migration guide for transitioning from previous versions due to these extensive architectural changes.