The CNCF Observability Summit North America 2026 opens in Minneapolis on May 21-22, bringing together practitioners and engineers across more than 45 sessions. This year's program includes a dedicated track exploring how AI and the Model Context Protocol are being applied in observability workflows - covering root cause analysis, tracing model-driven decisions, and incident response automation.

The framing from CNCF Executive Director Jonathan Bryce sets the tone: "AI is raising the stakes for reliability. As most organisations begin to run AI workloads, the hard question is whether they can trust and measure what's running in production. Observability is how the community closes that gap, and this summit is where that work happens."

The signal is clear. Observability is no longer just about collecting telemetry. The conversation has shifted to what you do with it when something breaks - how quickly you can reason across signals, identify root cause, and reduce the time an engineer spends assembling a picture from five different dashboards under pressure.

Alongside the AI + MCP track, the program addresses OpenTelemetry standards and the OpenTelemetry Transformation Language (OTTL), multi-cluster Kubernetes environments at scale, and real-world implementations including lessons on scaling OpenTelemetry, reducing costs, and debugging cloud failures. One session from AWS - "How Observability-First Development Lets You Ship Agents in Weeks, Not Months" - captures where the conversation is heading: observability as a prerequisite for building AI systems, not just operating them.

For Kubernetes and SRE teams, this summit arrives at a practical moment. The tools for collecting data are mature. The gap is in the investigation layer - connecting observability data to confident answers, automatically, before an engineer even opens a terminal. That is where the next reliability gains are.

At OpsWorker, that gap is exactly what we are building for. If you are attending the summit and thinking about the same problems - automated investigation, reducing toil on on-call rotations, or what AI-native incident response actually looks like in a Kubernetes environment - we would be glad to connect.

 Book a demo: calendly.com/opsworker/discover-opsworker   |   opsworker.ai