Keynote

Datadog was founded on the idea that engineering teams need a clear, shared view of their systems. That mission led us to build a platform that connects metrics, traces, and logs, and it continues to guide how we evolve. In this session, Kevin Hu, Staff Product Manager of Data Observability, will show how our founding principles still shape the way we build today. He’ll share recent product innovations and how they extend beyond traditional observability to help teams operate more reliably and respond faster when issues arise.

As AI transforms how you build, secure, and operate your systems, visibility and context have never been more critical. Datadog sits at the heart of your production systems, continuously ingesting, correlating, and contextualizing data across the entire stack, with operational context. This unified view provides a powerful foundation for applying AI safely and effectively. He will highlight how Datadog’s products enable intelligent, agentic capabilities that help teams detect, respond and build with greater confidence.

Observability tools have traditionally been part of a reactive process—waiting for incidents, then investigating and remediating them. But as AI rapidly advances, so have our AI augmented tools. In this session, Kai Xin Tai, Product Manager for Bits AI SRE at Datadog, will discuss the evolution of Bits AI SRE tool, share the latest updates, and show how it can proactively improve your reliability.

LLM applications are evolving fast, but without the right evaluations, iteration often feels like guesswork. In this session, Datadog Engineering Manager Greg Svigruha will share why evaluating LLM applications is so challenging, the best practices leading teams rely on to get the best results, and how Datadog LLM Observability's new capabilities makes systematic evaluation and experimentation simple.

See how Grainger enhanced reliability across a wide array of engineering teams through a scorecard system, with Gold, Silver, and Bronze Observability Medal Standards. They addressed the challenge of scaling reliability culture without overwhelming SREs, detailing the transition from monolith to microservices and the evolution of observability practices. They were able to foster continuous improvement through iterative processes by bringing teams into alignment and identifying critical paths for improved reliability.

Shared infrastructure means the risk of a noisy neighbor: a tenant that can impact the health or performance of other customer systems running nearby. At Zendesk they had to write a new playbook to recognize, examine, and test solutions for performance issues in multi-tenant systems, handling sudden capacity or latency issues without skyrocketing costs. How? A mix of tagging, query optimization, shared ownership and focus on customer experience. Now you can learn new ways to stay ahead of cost overruns and unexpected performance dips.