KubeCon + CloudNativeCon Europe 2025: Reflections

KubeCon + CloudNativeCon Europe 2025, held in London from April 1–4, was truly a landmark event for the cloud native community.

With around ~13K attendees, the conference showcased the remarkable maturity and continued innovation within the Kubernetes ecosystem as it celebrated its 10th anniversary.

As someone deeply immersed in the cloud native space, I found the experience both enlightening and energizing: a perfect blend of technical depth, community connection, and forward-looking insights.

Key themes and emerging trends

AI-enabled platform engineering

AI agents are now considered essential components of platform engineering efforts, even though the industry is still determining exactly how they will be applied. AI workloads need robust telemetry, runtime controls, and a well-defined operational model, meaning that successful platform engineering implementations can be the foundations for effective AI usage across an organization.

Check out Kagent, a Kubernetes native framework for building AI agents that was recently open-sourced and is being donated to CNCF as a sandbox project. Organizations were discussing prioritization of AI agent use cases around runtime security, reliability and incident remediation that affect their whole engineering landscape.

Unified observability and AI-driven insights

Breaking down silos between metrics, logs, and traces was a major focus, with significant progress on creating standardized query languages for observability data. OpenTelemetry has emerged as the de facto standard for observability, with organizations adopting it to standardize instrumentation across their applications. It’s only a matter of time before OpenTelemetry achieves graduated status with the CNCF.

eBay’s experiments with LLMs to summarize and explain observability data was particularly fascinating, showing how they’re managing 15PB of logs daily across 4,600 microservices. However, as one speaker noted, “If we attempted to feed all this observability data into an LLM, we would exceed its context window, leading to hallucinations and inaccurate summaries.”

FinOps and cost optimization

As Kubernetes deployments scale, cost management has become increasingly important. The conference featured discussions on using Open Policy Agent (OPA) beyond security for cost optimization, enforcing policies for resource usage efficiency, integrating with tools like OpenCost for real-time cost monitoring, and implementing “shift left” approaches to cost management. In the future, expect to see cost insights become as standard as monitoring dashboards.

Looking forward

The conference provided a clear vision of where the cloud native ecosystem is heading: toward a world where previously separate concerns are addressed holistically. Platforms have the potential to enable unprecedented agility and control, where security is built in by default, and where AI augments human capabilities rather than replacing them.