KubeCon 2025 Europe, London
KubeCon 2025 Europe in London marked the 10th anniversary of the Cloud Native Computing Foundation (CNCF)! Before diving into some highlights from the event, it’s a great time to take a moment to reflect on how it all started.
It was back in 2015 during OSCON (Open Source Convention) when Craig McLuckie shared his vision of making cluster-based, large-scale cloud computing accessible to everyone. The idea was to build a reference architecture that could be shared across cloud providers—standardizing tooling and enabling reusability of knowledge, instead of retraining engineers on every proprietary stack used by big tech.
This vision aimed to create technology that could outlive individuals, companies, and even entire eras—making it a smart investment to learn and adopt. To ensure openness and community stewardship, they partnered with the Linux Foundation to launch the CNCF and donated the very first project: Kubernetes.
OSCON 2015 recording: https://www.youtube.com/watch?v=mPhjFYXoAD0&list=PL055Epbe6d5YhDchEvY3O4nIuSLYyrx7K&index=56)
Fast forward to today, and it’s clear the philosophy has paid off. With over 30 graduated projects and many more in the incubation and sandbox stages, CNCF projects now underpin the infrastructure of countless organizations. They enable teams to focus on delivering value (applications) instead of reinventing the stack for running cloud-native apps at scale.
Platform Engineer Day
The conference kicked off with co-located events on “Day 0.” A major theme I followed was Platform Engineer Day.
It makes sense this topic continues to gain traction. With so many CNCF projects out there, setting up a production-ready environment is no small task. Requirements vary greatly between organizations—a fintech won’t have the same needs as a retailer or a university.
Challenges like portability, feature parity with existing platforms, and the availability of in-house knowledge all impact how quickly teams can mature their platform. Some problems can’t be solved with open source alone, and that’s where decisions around insourcing vs. outsourcing come into play.
This track was filled with practical talks—from hiring approaches and team building to common pitfalls and best practices. Real-world stories showcased how organizations are combining OSS with platform engineering to improve developer experience and operational resilience.
Platform engineering was a thread throughout the entire conference, not just Day 0. Highlights include:
CNCF Platforms whitepaper: https://tag-app-delivery.cncf.io/whitepapers/platforms/
Platform Engineering Maturity Model (2023): https://tag-app-delivery.cncf.io/whitepapers/platform-eng-maturity-model/
New certification tracks from the Linux Foundation:
Cloud Native Platform Engineering Associate (CNPA): https://training.linuxfoundation.org/certification/certified-cloud-native-platform-engineering-associate-cnpa/
Cloud Native Platform Engineer (CNPE): Coming soon
This continued investment from the Platform Working Group reaffirms that platform engineering is here to stay. More info: https://tag-app-delivery.cncf.io/wgs/platforms/
Another co-located event worth mentioning: Cilium. I attended a deep dive into how Cilium’s multi-cluster mesh leverages the multi-cluster API to make a service appear local to the cluster where the app is running. Very cool stuff.
And honestly, there were many other compelling themes worth exploring!
Sidenote: Sponsored Talks and Funding
You’ll find several sponsored talks, especially during the keynotes—which makes sense, considering CNCF’s 2023 annual report (https://www.cncf.io/reports/cncf-annual-report-2023/) notes that most of their revenue comes from event sponsorships, followed by membership dues, registration fees and a small portion from training programs.
Day 1
https://www.cncf.io/blog/2025/04/02/kubecon-cloudnativecon-europe-2025-day-one-keynote-recap/
A standout talk came from Christine Yen, CEO & Co-founder of Honeycomb:
Into the Black Box: Observability in the Age of LLMs (https://kccnceu2025.sched.com/event/1txBR/keynote-into-the-black-box-observability-in-the-age-of-llms-christine-yen-ceo-and-cofounder-honeycomb)
She explored how observability must evolve to support LLM-based systems, which behave probabilistically—unlike traditional code, which is deterministic.
Deterministic vs. Probabilistic Systems:
Traditional software: Same input → same output.
LLMs: Same input → potentially different outputs, depending on context.
Classic Testing: Unit tests, integration, mocks
LLM Testing: Behavioral tests, golden sets, hallucination rate tracking
Observability: Logging, tracing, error rates
LLM Observability: Prompt performance, model drift, hallucination detection
Debugging: Stack traces vs. prompt tuning and dataset curation
In short, the parallels exist, but the methods differ—an inspiring look at the challenges ahead.
Another great session:
The Life (or Death) of a Kubernetes API Request, 2025 Edition by Abu Kashem (Red Hat) & Stefan Schimanski (Upbound) https://sched.co/1tx89
A deep dive into how Kubernetes handles API requests—shedding light on the internals and offering tips for troubleshooting complex cluster issues. It emphasizes the critical role of platform engineers in ensuring the reliability and efficiency of Kubernetes clusters. The presentation highlighted key aspects such as the complexities of managing Kubernetes, the importance of observability, and the challenges faced during production incidents. Through real-world anecdotes and technical insights, they approached the human side of cloud-native operations, offering valuable lessons for engineers navigating the Kubernetes ecosystem.
Day 2
https://www.cncf.io/blog/2025/04/03/kubecon-cloudnativecon-europe-2025-day-two-keynote-recap/
A notable observation was the OpenInfra Foundation logo passing by on one of the slides. Known for projects like OpenStack, focused on open-source infrastructure solutions. As it turns out they’ve announced plans to join the Linux Foundation.
One highlight talk:
From Metal To Apps: LinkedIn’s Kubernetes-based Compute Platform by Ahmet Alp Balkan & Ronak Nathani
LinkedIn operates over 500,000 servers across multiple data centers, hosting more than 3,000 services and handling over 1.5 million containers with 50,000+ deployments daily. To manage this scale, they've developed a custom Kubernetes-based compute stack comprising several key components:
Compute Broker: A gRPC API that manages machine pools and allocates resources without relying on a Kubernetes distro (but rather an inhouse setup)
Host Health Monitoring & Remediation: Automated detection and remediation of unhealthy hosts, ensuring minimal manual intervention
Maintenance Orchestrator: Facilitates rolling updates across the fleet, including OS and kubelet upgrades, without disrupting services
The platform supports various workloads, including stateless applications, stateful services, and batch jobs, all while maintaining tenant isolation and resilience. By leveraging Kubernetes as a foundational primitive, LinkedIn has achieved a scalable and flexible compute platform capable of handling diverse workloads efficiently.
While they are still migrating all services to Kubernetes these are some valuable lessons they stated:
- Start early and make incremental progress
- Figure out which tech debt to solve now vs later
- Be intentional about what features to use in Kubernetes
- Don’t give raw Kubernetes to your customers
- Invest in building abstractions
- Invest in guard-rails to prevent user errors
- Develop good user guides for self-serve troubleshooting
Day 3
https://www.cncf.io/blog/2025/04/07/kubecon-cloudnativecon-europe-2025-day-three-keynote-recap/
Unfortunately, I couldn’t attend Day 3 live, so I missed the sessions. Looking forward to watching the recordings.
Closing Note
All in all, the event had a fantastic vibe. The venue could’ve been better—but in the end, it’s the people that make it worthwhile. Always inspiring to be around a community pushing the cloud-native ecosystem forward.
Looking Ahead
As CNCF celebrates a decade of shaping the cloud-native world, it’s clear the journey is far from over. With new certification paths, ongoing innovation in platform engineering, and open-source projects maturing at scale, the next 10 years are bound to bring even more transformation—especially as AI, sovereignty, and sustainability take center stage.
Let’s Connect
Were you at KubeCon EU 2025? Or have thoughts on the future of platform engineering and cloud-native tooling? I’d love to hear what stood out to you—feel free to reach out!
