KubeCon + CloudNativeCon EU 2026 in Amsterdam will bring the cloud native community together with projects, vendors and end users in what is one of the largest open source events.
In this post, we highlight sessions and activities that may be of particular interest to end users, including dedicated discussion forums and recommendations from members of the CNCF End User Technical Advisory Board (TAB) along with practical tips for making the most of the event.
End User content
The TAB has thought of different ways where the experience of end users at KubeCon + CloudNativecan be enhanced This will be a gradual process, but some of the ideas have been implemented in advance of Kubecon + CloudNativeCon Europe in Amsterdam.
Make sure you do not miss these opportunities to discuss with other end users.
Birds of Feather (BoF)
We will have three BoF sessions targeting end users, as a space to focus and discuss how we’re targeting specific topics and come out with best practices and gaps to be tackled. For this Kubecon + CloudNativeCon we’ll be covering:
- BoF: AI Infrastructure and Platform: LLMs, Agentic AI
- Convener: Joseph Sandoval
- Best practices for generic MLOps on cloud native platforms, common pain points of running such workloads on cloud native infra, gaps in the existing ecosystem
- BoF: Infrastructure Optimization for GPUs / Inference / Training / Networking
- Convener: Ricardo Rocha
- Low level runtime optimizations, CPU/GPU NUMA affinity and pinning, low latency networking, support for heterogeneous hardware
- BoF: AI Observability
- Convener: Alolita Sharma
- Support for observability of inference, training and agent workloads, tools being used today in production, gaps in the existing ecosystem
Town Hall
We’ll have our usual town hall where end users can come forward with questions to the TAB. But this time we change the format, starting the session with a series of short 5min presentations that should trigger discussion between the TAB and the audience.
Poster sessions
The effort continues on publishing more reference architectures from our end users, you can check them here. We’re giving the possibility for owners of published architectures to have a poster during the poster session to showcase their own publications.
Pass by and you can directly ask your questions.
End User reception
On Tuesday 17:30 – 18:30 there will be a reception for CNCF end user members. Join us at the Grand Cafe, Ground Level – Entrance K side of the building.
Session highlights
Below you can find session recommendations and additional tips on how to make the best of your KubeCon + CloudNativeConexperience this year, from the members of the TAB.
Don’t hesitate to reach out and grab us for a chat in the hallway!
Ahmed Bebars
1000 Services, 1 Year, 0 Downtime: Airbnb’s Zonal Cluster Migration
Speakers: Sunny Beatteay, Airbnb
Why I want to see it: I’m always looking for real stories of large organizations moving critical workloads without reader‑visible downtime. Airbnb talks through how they migrated around a thousand services to a zonal topology in a year with zero downtime. I’m particularly interested in the migration strategies, blast‑radius thinking, and what they’d do differently next time.
Kubernetes Autopsy: Live Debugging a Cluster Meltdown (Disney+ Hotstar)
Speakers: Sandeep Kanabar (Gen), Aditi Gupta (Disney+ Hotstar), Anshika Gupta (Independent)
Why I want to see it: News and media platforms experience traffic spikes and hard SLOs similar to streaming. I’m keen to hear how Disney+ Hotstar approached a real Kubernetes cluster meltdown: what their telemetry showed (or failed to show), how they debugged under pressure, and what permanent changes they made to avoid a repeat.
Smart Routing at Scale: How Spotify’s XDS Control Plane Cut 75% of Cross‑Zone Traffic
Speakers: Yannick Epstein, Anya Hristova – Spotify
Why I want to see it: Multi‑AZ and multi‑region setups easily leak money through unnecessary cross‑zone traffic. Spotify’s talk on how they used an XDS control plane to cut cross‑zone traffic by 75% is exactly the kind of deep technical routing and control‑plane story that can translate into both cost savings and better latency for users.
We Deleted Our Observability Stack and Rebuilt It With OTEL: 12 Engineers to 4 at 20K+ Clusters (DigitalOcean)
Speakers: Yash Sharma, Kunju Perath – DigitalOcean
Why I want to see it: Observability is often where cost, complexity, and toil quietly accumulate. DigitalOcean’s story of rebuilding their observability stack around OpenTelemetry for 20K+ clusters—while shrinking the team from 12 to 4—sounds like a masterclass in focusing on the right signals, rethinking architecture, and balancing visibility with cost.
Alolita Sharma
Collisions in the Dark: Illuminating the 95% of Kubeflow You Can’t See
Speakers: Amine Lahouel and Laura Llinares, CERN
Why I want to see it: This is an interesting session where CERN scientists will be discussing what they need from an observability perspective on the performance of their ML platform infrastructure. Gaining deep, actionable insights into the performance, resource consumption, and overall health of the entire system are major KPIs for end-users. Understanding CERN’s takeaways and challenges in measuring ML platform performance with these KPIs will be valuable.
Discover Cortex: High Scalability Metrics in 2026
Speakers: Friedrich Gonzalez and Charlie Le, Apple
Why I want to see it: This session aims to dive into the latest features in Cortex, a multi-tenant framework for observability metrics storage, as well as the project roadmap on what’s rolling out next for scalability and reliability.
Is the Agent in the Room with Us Right Now?
Speakers: Nick Rutigliano and Andrew Halaney, Netflix
Why I want to see it: At Netflix scale, moving traditional workloads to agentic AI requires understanding pros and cons of isolation. These requirements apply to security, RBAC network configurations, containers, and user namespaces, and general noisy neighbor improvements and can be useful takeaways for other end-users.
GenAI Observability: Keeping GenAI Honest Without Oversharing
Speaker: Liudmila Molkova, GrafanaLabs
Why I want to see it: Most end-users are leveraging Generative AI for their solutions and services. This talk will focus on how to instrument your AI applications with OpenTelemetry following the Generative AI Semantic Conventions. Understanding performance and usage telemetry data is important but understanding how to handle separate data for compliance and cost information is also valuable.
Tips for the best experience
- Select topics and tracks that are most relevant for you. I’ve found this helps me short-list the sessions I’m most interested in
- Plan for the time needed to get from one session to the next
- Hallway discussions are very valuable
- Drop by the project pavilion. If you have a list of favorite projects, the project pavilion is a great place to meet the project maintainers in person
- Have fun and enjoy learning from the conference sessions
Ben Somogyi
Enterprise Challenges with MCP Adoption
Speakers: Christian Posta (Global Field CTO at Solo.io)
Why I want to see it: In this session, Christion will be discussing the Model Context Protocol (MCP) and discussing some of the issues surrounding the establishment of MCP servers that provide access to enterprise APIs on behalf of a user. Understanding how other companies are addressing these problems to prevent misuse and to protect their business through governance will be helpful to understand some best practices and strategies to leverage AI securely within your use case.
Navigating the Gateway API Maze: 40+ Implementations, 55+ Features, and a Path to Portability
Speakers: Cristine Kim (Software Engineer, Isovalent at Cisco) and Beka Modebadze (Software Engineer @ Google)
Why I want to see it: In this talk Christine and Beka plan to “cut through the confusion and provide you with the knowledge to navigate the Gateway API landscape”. In a dynamic space where capabilities like the nginx ingress controller have been recently retired this talk will help provide some insights of how to navigate the Gateway API space and discover implementations that will be portable and remain vendor agnostic, helping you build a more resilient platform with respect to k8s networking.
Kenta Tada
10 Years of Cilium: Connecting, Securing, and Simplifying the Cloud Native Stack
Speakers: Bill Mulligan and Paul Arah, Isovalent at Cisco; Marcelo Mello, Celonis; Neha Aggarwal, Microsoft
Why I want to see it: Cilium has become a crucial piece of software supporting modern container networking and security. In this session, they will be sharing the latest updates alongside real-world stories from our users. It’s a great chance to see how it can simplify and secure your own infrastructure.
Ricardo Rocha
Gateway API: Bridging the Gap from Ingress to the Future
Speakers: Nick Young and James Strong, Isovalent at Cisco; Katarzyna Łach and Rostislav Bobrovsky, Google; Norwin Schnyder, Airlock
Why I want to see it: the transition from Ingress and services to the Gateway API is not trivial but it’s an opportunity for all end users to improve and simplify the ways they expose their endpoints. Listening from experts should bring a lot of good information.
Virtual Power Plants (VPP): How They Work and What They Are
Speakers: LeRenzo Malcom and Mario Flores, Enpal
Why I want to see it: New use cases for cloud native usually make great talks. Managing power plants and solar devices as cloud native resources covers distributed, edge use cases and should bring useful information to all of us.
Instrumenting Kueue Scheduling for ML Training
Speakers: Amy Chen, CoreWeave and Gabriel Saba, Google
Why I want to see it: Kueue is an important tool for advanced scheduling and other capabilities needed for large scale, distributed ML training. A large end user’s feedback on their experiences along with one of the Kueue maintainers should make a great session.
The Shell Awakens: Cloud Native Workflows for Particle Physicists
Speakers: Raulian-Ionut Chiorescu and Hannes Hansen, CERN
Why I want to see it: This one comes from our team. Raul and Hannes will describe our next-gen scientific computing platform, and how we made the old and new ways of working compatible and used on top of the same resources and stack.
Tips for the best experience
- Kubecon + CloudNativeCon has lots of sessions in a very large venue. Don’t overcommit the sessions you want to attend, consider the time to walk between rooms and leave time to ask questions to speakers after the sessions end
- Spend time in the hallway, reach out to people, network as much as possible. These are the people that will help you out once back in the office, debugging or building new stacks. Also… check out the multiple parties happening every evening
- Note down highlights and the names of those that can help you dig further later on. Bring it back to your teams, dedicate time to present and discuss your findings back home
- Have fun! There is a lot of content and knowledge on display, but above all this event showcases and supports the community that keeps cloud native going strong!