Grafana Labs has published major updates across two of its core observability products: Grafana 12.3, and Grafana Tempo 2.9. The two releases have distinct improvements in monitoring and tracing for Grafana users.
Grafana 12.3 adds new visual tools and learning experiences to its unified dashboard environment. The release includes a fully rebuilt logs panel, now with colour highlighting, flexible client-side search and filtering, and support for both millisecond and nanosecond precision in timestamps. It also introduces a “logs context” feature, allowing users to view events that occurred immediately before or after a selected log line, with time ranges adjustable from roughly one hundred milliseconds up to two hours. There is also a field selector component added; this shows which log fields appear most often, lets users toggle those visible fields, reorder them and reduce noise. And there is a new “Interactive Learning” feature. The blog explains that “Interactive Learning provides context-aware guidance based on where you are and what you’re doing in Grafana.” These changes aim to make logs exploration faster and more intuitive.
Grafana Tempo 2.9 enhances tracing, and the biggest addition in this release is experimental Model Context Protocol (MCP) server support, which allows AI assistants to access tracing data with TraceQL queries. It adds probabilistic sampling hints in TraceQL, for example with(sample=true) or with(sample=0.xx), which allow users to sacrifice some accuracy for speed in high-volume environments. The release also improves operational visibility in multi-tenant setups thanks to new metrics such as bytes inspected during queries and span timestamp distance, both into the past and the future.
In comparison, several existing open source tools have offered parts of this functionality but not the same combinations. Tempo’s new tightly integrated sampling control and AI-assistant features brings it in line with other OpenTelemetry-derived tracing tools. Regarding Grafana 12.3, community feedback highlights appreciation for more powerful logs visualisations. Some suggest that similar features are found in commercial log tools like Datadog or Splunk, though those often come at a higher cost and without native integration into open source stacks. In tracing, Tempo 2.9’s new sampling hints have been well received from those who work with high trace volumes. In a LinkedIn post, Deutsche Bank’s Florin Lungu lauded the new ability to make better decisions in performance monitoring thanks to the new MCP server support.
I found it interesting that this integration significantly enhances our ability to glean insights from complex data.
Grafana Labs have advised caution to those looking to upgrade. The company advises careful planning and testing before upgrading. To summarise, Grafana 12.3 advances log exploration and usability. Grafana Tempo 2.9 has better trace sampling and introduces AI-driven contextual access. Both of these products are available now.
