The Road To KubeCon EU 2025: Top 10 Must-Attend Sessions 

KubeCon week is upon us, and everyone in the community is excited to flock to London for another week of knowledge exchange, networking, and celebration of our nerdiness…and swag! Let us not forget about swag! 

Team Komodor is on the ground, ready at booth #N330 to demonstrate our latest advancements with Drift Management and freeform chat with KlaudiaAI. We’ve also got awesome transformer-inspired swag to give away, a Valve Steam Deck, rare Lego sets, and of course – another iconic Kaptain K8s shield! So, come geek out with me and the team, and see the future of AI-powered Kubernetes operations in action!  

But what about the agenda? 

KubeCon is packed with cutting-edge content, especially for seasoned Kubernetes practitioners. With hundreds of sessions spanning operations, observability, platform engineering, and AI-driven automation (AIOps), it can be challenging to decide where to invest your time. To help you optimize your schedule, we’ve reviewed the full agenda (Schedule | LF Events) and hand-picked ten must-attend sessions that promise deep insights for advanced users. From real-world case studies of running Kubernetes at scale to the latest in observability tooling and AI integrations, these talks will equip you with the knowledge to level up your K8s game. 

Let’s dive in and see what KubeCon + CloudNativeCon Europe 2025 has to offer!

1. Superpowers for Humans of Kubernetes: How K8sGPT Is Transforming Enterprise Ops

Speakers: Alex Jones (Principal Engineer, AWS) & Anaïs Urlichs (Platform Engineer, JP Morgan Chase) – When: Wednesday, April 2, 2025, 14:30–15:00 BST (KubeCon + CloudNativeCon Europe 2025: Superpowers for Humans of Kubernetes: Ho…). This breakout session explores how the open-source K8sGPT project uses AI to streamline Kubernetes troubleshooting and operations at scale (KubeCon + CloudNativeCon Europe 2025: Superpowers for Humans of Kubernetes: Ho…). The speakers will demonstrate how AI-driven diagnostics can identify and explain complex cluster issues before they impact users, effectively augmenting human operators. It’s highly relevant to platform SREs and ops teams looking to do more with less – you’ll learn how to leverage AI to triage incidents faster and manage dozens of clusters with a lean team, turning tedious toil into automated insight.

2. Prometheus Deep Dive: What’s New in v3.0 and Beyond

Speakers: Saswata Mukherjee (Senior Software Engineer, Red Hat) & Fiona Liao (Staff Software Engineer, Grafana Labs) – When: Wednesday, April 2, 2025, 14:30–15:00 BST (KubeCon + CloudNativeCon Europe 2025: Prometheus Deep Dive: What’s New in v3.0…). If you’re responsible for monitoring Kubernetes, this talk is a don’t-miss update on the Prometheus ecosystem. Two Prometheus maintainers will walk through the major new features and enhancements introduced in Prometheus 3.0 (including a refreshed UI/UX) and what’s coming next (KubeCon + CloudNativeCon Europe 2025: Prometheus Deep Dive: What’s New in v3.0…). Expect a deep dive into how these improvements make monitoring more efficient and scalable. Advanced users will gain practical tips on upgrading to v3, leveraging its new capabilities, and even influencing upstream development with their feedback.

3. Taming 50 Billion Time Series: Operating Global-Scale Prometheus Deployments on Kubernetes

Speakers: Orcun Berkem & Alan Protasio (AWS) – When: Wednesday, April 2, 2025, 12:00–12:30 BST (KubeCon + CloudNativeCon Europe 2025: Taming 50 Billion Time Series: Operating…). Operating Prometheus for massive scale metrics is a true engineering challenge – and this session tackles exactly that scenario. AWS engineers share how they scaled Prometheus to handle 50 billion active time series across 20 regions on Kubernetes (KubeCon + CloudNativeCon Europe 2025: Taming 50 Billion Time Series: Operating…). They will detail the architecture (like using stateful sets and cell-based sharding), deployment processes for high availability, and multi-tenancy safeguards (shuffle sharding, token bucket rate limiting) that keep such a large system reliable. Advanced practitioners in charge of large clusters or multi-tenant platforms will learn battle-tested techniques for building resilient and efficient monitoring systems at global scale (KubeCon + CloudNativeCon Europe 2025: Taming 50 Billion Time Series: Operating…).

4. Keynote: Into the Black Box: Observability in the Age of LLMs

Speaker: Christine Yen (CEO & Cofounder, Honeycomb) – When: Wednesday, April 2, 2025, 09:11–09:26 BST (KubeCon + CloudNativeCon Europe 2025: Keynote: Into the Black Box: Observabili…). In this insightful keynote, observability expert Christine Yen addresses the challenges of operating applications that incorporate large language models (LLMs). She argues that deploying AI/ML features is “just like building on any other black box” and advocates for applying the same observability best practices (like SLOs and high-cardinality telemetry) to these unpredictable, probabilistic systems (KubeCon + CloudNativeCon Europe 2025: Keynote: Into the Black Box: Observabili…). Advanced Kubernetes users will appreciate the nuanced discussion of performance and reliability issues that arise with AI-driven services. Expect to come away with concrete approaches to monitor and troubleshoot LLM-powered apps in production – making this highly relevant as AI workloads become more common in cloud-native environments.

5. Keynote: The Observability Platform Engineering Advantage: From Zero-Code to Monitoring as Code

Speaker: Kasper Borg Nissen (Developer Relations Engineer, Dash0) – When: Wednesday, April 2, 2025, 10:17–10:32 BST (KubeCon + CloudNativeCon Europe 2025: Keynote: The Observability Platform Engi…). This keynote sits at the intersection of observability and platform engineering – a hot area for advanced practitioners. Kasper Nissen will outline how platform teams can bake observability into their infrastructure from the ground up using open standards. He’ll showcase the power of OpenTelemetry for unified tracing/metrics/logs and how “zero-code” instrumentation can capture app signals automatically (KubeCon + CloudNativeCon Europe 2025: Keynote: The Observability Platform Engi…). More importantly, you’ll learn how treating monitoring as code enables consistent, repeatable observability across all your clusters and apps. If you’re building an internal platform or managing Kubernetes at scale, this talk will provide a vision (and examples) of bridging dev, ops, and monitoring efforts into one cohesive strategy (KubeCon + CloudNativeCon Europe 2025: Keynote: The Observability Platform Engi…).

6. The Bricks That Make Us – How The LEGO Group Avoids 50 Mediocre Kubernetes Implementations

Speakers: Thomas Øther Rasmussen & Paul Farver (The LEGO Group) – When: Wednesday, April 2, 2025, 15:15–15:45 BST (KubeCon + CloudNativeCon Europe 2025: The Bricks That Make Us – How the LEGO G…). Running Kubernetes in an enterprise with many teams can feel like herding cats – and LEGO’s platform engineering team has tackled this challenge head on. In this talk, they discuss how they balance developer autonomy with governance to prevent every team from running a separate “mediocre” k8s stack (KubeCon + CloudNativeCon Europe 2025: The Bricks That Make Us – How the LEGO G…). You’ll hear how 100+ product teams at LEGO are onboarded onto a central container platform without stifling creativity (no “Kubernetes Police” here!). For advanced platform engineers, this session offers valuable lessons on fostering good communication between infrastructure, platform, and application teams, and keeping developers happy while maintaining consistency. It’s a rare peek into how a global company avoids chaos and delivers a great developer experience on Kubernetes.

7. Day-2’000 – Migration From Kubeadm+Ansible to ClusterAPI+Talos: A Swiss Bank’s Journey

Speaker: Clément Nussbaumer (PostFinance) – When: Wednesday, April 2, 2025, 12:00–12:30 BST (KubeCon + CloudNativeCon Europe 2025: Day-2’000 – Migration From Kubeadm+Ansib…). Migrating the core of your Kubernetes platform is a risky endeavor, especially in a regulated, air-gapped bank environment. In this technical session, PostFinance shares how they moved 35 clusters from a legacy kubeadm/Ansible/Puppet setup to a modern ClusterAPI+Talos stack with zero downtime (KubeCon + CloudNativeCon Europe 2025: Day-2’000 – Migration From Kubeadm+Ansib…). The talk will delve into the migration process, including tooling built to automate the transition and thorny issues encountered (like etcd quorum loss, reconciling API server configs, and encryption key mismatches between old and new clusters (KubeCon + CloudNativeCon Europe 2025: Day-2’000 – Migration From Kubeadm+Ansib…)). A live demo will illustrate their approach to fleet management using ArgoCD and Talos config files. This is a must-see for Kubernetes operators planning major upgrades or looking to streamline day-2 operations – you’ll gain real-world insights on executing large-scale changes in production without breaking everything.

8. Keynote: Driving Innovation at Michelin: How We Scaled Cloud & On-Prem Infrastructure While Cutting Costs

Speakers: Gabriel Quennesson & Arnaud Pons (Michelin) – When: Thursday, April 3, 2025, 09:25–09:30 BST (KubeCon + CloudNativeCon Europe 2025: Keynote: Driving Innovation at Michelin:…). In just five minutes, this lightning keynote from Michelin manages to pack in an impressive story of enterprise Kubernetes transformation. The speakers will explain how Michelin re-architected its Kubernetes platform to support 441 applications across 62 clusters spanning cloud and on-prem environments (KubeCon + CloudNativeCon Europe 2025: Keynote: Driving Innovation at Michelin:…). They achieved this by adopting Cluster API for automated cluster management, Crossplane for cloud provisioning, and GitOps with ArgoCD – resulting in a 44% reduction in platform costs and 85% faster upgrade cycles (KubeCon + CloudNativeCon Europe 2025: Keynote: Driving Innovation at Michelin:…). For advanced practitioners, the value is hearing how a large, traditional company optimized its Kubernetes operations at scale. You’ll takeaway strategies for cutting costs and complexity in your own environments, and proof that even conservative industries can embrace cloud-native automation to drive efficiency (and even help with talent retention by using modern tooling!).

9. Keynote: LLM-Aware Load Balancing in Kubernetes: A New Era of Efficiency

Speakers: Clayton Coleman (Distinguished Engineer, Google) & Jiaxin Shan (Software Engineer, ByteDance) – When: Friday, April 4, 2025, 09:06–09:21 BST (KubeCon + CloudNativeCon Europe 2025: Keynote: LLM-Aware Load Balancing in Kub…). As large language model applications proliferate, Kubernetes developers are finding that traditional load balancing strategies (like round-robin or simple QPS-based routing) fall short for LLM workloads (KubeCon + CloudNativeCon Europe 2025: Keynote: LLM-Aware Load Balancing in Kub…). In this forward-looking keynote, industry heavyweight Clayton Coleman and Jiaxin Shan introduce new Kubernetes APIs designed for LLM-aware traffic routing. They’ll describe how LLM requests vary wildly in CPU/GPU usage and duration, and present a solution that lets you define serving objectives and priorities for AI inference jobs. The proposed APIs integrate with the Gateway API, making it easy to plug into existing ingress controllers for smart request scheduling (KubeCon + CloudNativeCon Europe 2025: Keynote: LLM-Aware Load Balancing in Kub…). Advanced attendees will get a glimpse of Kubernetes’ future in the AI era – including a demo of how these enhancements drastically improve utilization and response times for real-world AI workloads. If you’re interested in Kubernetes + AI (a key part of “AIOps”), put this on your schedule.

10. Expanding eBPF’s Reach: From Batteries-Included Auto-Instrumentation to E2E Observability Pipelines

Speaker: Dom Del Nano (Cosmic) – When: Wednesday, April 2, 2025, 12:00–12:30 BST (KubeCon + CloudNativeCon Europe 2025: Expanding eBPF’s Reach: From Batteries-I…). This talk is all about pushing the boundaries of eBPF for observability. eBPF-based auto-instrumentation (as seen in tools like Pixie) offers magical visibility without changing code, but it isn’t always flexible enough for every scenario (KubeCon + CloudNativeCon Europe 2025: Expanding eBPF’s Reach: From Batteries-I…). Dom Del Nano will discuss a “batteries included but removable” approach – essentially using eBPF for out-of-the-box data collection while still allowing engineers to customize and extend their telemetry when needed. He’ll highlight how CNCF projects Pixie and Inspektor Gadget can be combined to unlock eBPF’s full potential, turning Pixie’s data collector into a universal agent powering tailored observability pipelines (KubeCon + CloudNativeCon Europe 2025: Expanding eBPF’s Reach: From Batteries-I…). For seasoned observability engineers, this session promises insight into evolving past one-size-fits-all monitoring. You’ll learn how to get deep visibility (system calls, kernel events, etc.) with eBPF and then enrich or filter that data to suit your platform’s unique needs – achieving end-to-end observability that’s both automated and adaptable.

Final Note

KubeCon + CloudNativeCon Europe 2025 has no shortage of advanced content, but the sessions above stand out for anyone seeking to refine their Kubernetes operations and platform expertise. By focusing on these talks, you’ll hear from experts and practitioners who are solving the same challenges you face – whether it’s scaling clusters to new heights, integrating AI into ops, building internal platforms, or mastering observability tooling. Be sure to plan ahead (some of these run in parallel), and take advantage of the conference’s recording availability – all keynotes and sessions will be recorded and posted on the CNCF YouTube channel within two weeks (Schedule | LF Events). With an agenda this rich, a bit of strategy will help you maximize learning and come back from KubeCon + CloudNativeCon with actionable insights to apply in your own Kubernetes journey. Enjoy the conference!