Agents in Production: The Future of AI-Driven Observability

In this episode of Grafana’s Big Tent, host Tom Wilkie is joined by Spiros Xanthos (Founder & CEO, Resolve AI), Manoj Acharya (VP of Engineering at Grafana Labs), and Cyril Tovena (Principal Software Engineer, Grafana Assistant team) to explore the rapidly evolving world of agentic AI in production systems. They dig into automated root cause analysis, knowledge graphs, and how AI agents are beginning to operate alongside — and sometimes ahead of — human engineers. The conversation tackles trust, pricing models, data privacy, and whether we’re approaching a future where incidents resolve themselves. Plus: Tesla autopilot analogies, AI-written performance reviews, and what’s next beyond observability.

In this episode, hosts Mat Ryer and Matt Toback sit down with Ted Young, co-founder of OpenTelemetry, to unpack the real story behind OTel — from its Google origins to today’s push toward zero-touch observability. They explore how OpenTelemetry emerged from the merger of OpenTracing and OpenCensus, why the “three pillars” model may be holding teams back, and what the infamous logs vs. events debate reveals about how the industry thinks about data. Ted breaks down the technical and philosophical shifts happening in observability, including context propagation, structured telemetry, semantic conventions, and why backwards compatibility matters more than most people realize.

Mat Ryer and Tom Wilkie sit down with Grafana creator, Torkel Ödegaard, and Grafana Labs' Distinguished Engineer, Ryan McKinley, to unpack Grafana 12’s biggest shifts and experiments. Topics include the new dashboard API and Git sync workflows for dashboards as code, dynamic/constraint-based layouts and conditional panels, SQL expressions for joining query results, and the new assistant sidebar. They also reflect on Grafana’s history,  how ease of use and looking “cool” helped adoption, and practical guidance for teams operating at scale.

Hosts Mat Ryer and Matt Toback welcome Grafana engineers Sven Großmann, Alexa Vargas Ortega, and Lukasz Gut for a deep dive into Grafana’s internal hackathons. They discuss ideation, team formation, tooling (including AI-assisted coding), and the path from hackathon prototype to product feature — plus the occasional bird- and board-game tangent. 

In this episode of Grafana’s Big Tent, hosts Mat Ryer and Tom Wilkie sit down with Marco Pracucci (Principal Software Engineer, Grafana Labs), Cyril Tovena (Principal Software Engineer, Grafana Labs, and Ryan Worl (Co-founder, WarpStream) to unpack the next-generation architecture for Grafana Mimir. They discuss why decoupling the read and write paths matters, how Kafka-style partition guarantees change the game, and why replication-factor-three had become a costly bottleneck. Along the way: multi-AZ realities, object storage “loopholes,” and what new flexibility unlocks across the Grafana stack.

In this episode of Grafana’s Big Tent, hosts Mat Ryer and Tom Wilkie sit down with Julius Volz, co-creator of Prometheus and founder of PromLabs, and Richi "RichiH" Hartmann, Prometheus maintainer and Senior Developer Programs Director at Grafana Labs, to unpack Prometheus 3.0. They trace Prometheus back to its beginnings at SoundCloud, where Julius and his colleagues were trying to recreate the magic of Google’s Borgmon. The group dives into why Prometheus stuck with a text-based exposition format, what’s new in 3.0 — including performance gains, revamped UI, UTF-8 support, and native histograms — and how it fits alongside OpenTelemetry. Along the way, they explore governance changes, future directions, and why emojis are now valid in metric names.

Recorded live at GrafanaCon Seattle, hosts Mat Ryer and Tom Wilkie kick off Season 3 of Grafana’s Big Tent by welcoming Ivana Huckova (Staff Software Engineer, Grafana Labs), Andrew McCalip (Head of Research and Development, Varda Space Industries), and Brad Fitzpatrick (Chief Engineer, Tailscale). They swap first-tinkering stories and laugh through IoT wins and fails (dog happiness monitors, candle sensors—oops). Andrew explains how his self-righting, foam-filled autonomous boat survives the open seas, while Brad shares how he uses AI to 3D-print OpenSCAD parts on demand.

In the season two finale, hosts Tom Wilkie and Mat Ryer are joined by Richard "RichiH" Hartmann, Director of Community at Grafana Labs, and Pavel Strobl of Prusa Research to explore the intersection of 3D printing and observability. Prusa Research not only produces 3D printers that both Tom and Richi use to create everything from wind tunnels to miniature Grots, but the company also uses Grafana internally for their own systems, making this a dynamic conversation about how Grafana is being used to monitor and improve physical manufacturing processes as well as IT infrastructure.

Two members of the OpenTelemetry governing board, Daniel Gomez Blanco and Juraci Paixão Kröhling, join Mat and Matt to dive deep into the evolution of OTel from its early days (including the story of how OTel got its name) to its current state and future roadmap. 

Hosted by Tom Wilkie, this week's episode features Thomas Dullien, former Co-Founder and CEO of Optimyze; Bryan Boreham, Grafana Labs' distinguished engineer; and Senior Software Engineer at Grafana Labs and member of the CNCF Environmental Sustainability TAG Niki Manoledaki. Together, they delve into the complex world of cloud resource management and sustainability and explore the intricate challenges of optimizing Kubernetes clusters, from understanding allocation versus utilization to developing strategies that reduce waste and carbon footprint.