On a bright sunny London day, attendees from all over the world flocked to the ExCel Arena for the first day of KubeCon + CloudNativeCon 2025. Here’s a quick recap of the keynotes.

The crowd was welcomed by Cloud Native Computing Foundation CTO and founder Chris Aniszczyk, who talked about the tenth anniversary of cloud native computing and walked everyone down memory lane. Chris updated the audience on a number of efforts – from Golden Kubestronauts to training efforts in Africa (a partnership with Andela) – and also talked about the beta release of GitJobs.dev, a way to find open source jobs. Then Vasu Chandrasekhara from SAP joined him to explain NeoNephos, a sovereign cloud-edge continuum for Europe.

Christine Yen, CEO and cofounder of Honeycomb, presented a keynote on Observability in the age of LLMs. Christine explained that writing software today feels more like magic than ever, and that while LLMs are like APIs, they complicate the traditional ways of doing things because they can be difficult to test, mock up and debug. The answer, she argued, was observability, because it helps embrace the unpredictability that LLMs bring. The trick is to bundle observability with evals so teams can capture the good and bad as it happens and create systems that evolve with the code. The good news? This isn’t a new skillset. A lot of this shift is already underway.

Vijay Samuel, principal MTS and architect at eBay, spoke about AI-enabled observability explainers – We actually did something with AI! With 4600 microservices powering the eBay site, and 15PB of logs per day, Vijay explained how incidents impacted the team’s ability to provide the highest quality customer experience and that trying to deal with all this data manually is cumbersome. So his group started small experiments using LLMs as observability building blocks that could summarize and explain critical paths. They found simple flows they could automate and in the end leaned into LLMs’ strengths all the while realizing there were still things best done with code instead.

Greg Kroah-Hartman, Linux Kernel Maintainer and Fellow with The Linux Foundation, spoke on Rust in the Linux Kernel: A new era for cloud native performance and security. Greg walked the audience through the close relationship between Rust and Linux and shared some statistics, including the fact that Linux has 4,992 developers and 355+ employers. 

Rob Koch, principal of Slalom Build and an American Sign Language user, shared his experience working toward machine recognition of sign language, in Empowering accessibility through Kubernetes: The future of real-time sign language interpretation. Currently only two  or three companies in the world are working on this, Rob explained, and his company is using Kubeflow and Kubernetes to tackle the challenge. Kubeflow simplifies the deployment of machine learning models, ensuring smooth transitions from training to production.In the future, the company wants to prepare for recognizing other sign languages and seeing things from other viewpoints.

Kasper Borg Nissen, developer relations engineer at Dash0, spoke on The observability platform engineering advantage: from zero code to monitoring as code. His contention is that current systems are really inefficient when it comes to observability. “We don’t have a metrics problem, or a tracing problem. We have systems problems. And yet, many of us still treat these as separate entities. We have one browser tab for logs, another for metrics, and a third for tracing. We are relying on humans to correlate signals across them. It’s inefficient, it’s error-prone, and frankly, it’s not how modern observability should work.”

He ended with a call to action to apply platform engineering principles to observability: “Applying platform engineering principles can transform observability from an afterthought into a seamless, scalable, and developer-friendly experience.”

Get more of Day One in this video.