CNCF-hosted Co-located Events Europe 2025

Over the past 8 years, service meshes have provided the Kubernetes and CNCF ecosystems with the networking capabilities necessary for a robust microservice architecture. However, Windows container users have often been unable to take advantage of these features because of the deep Linux dependencies. Until now.

In this talk, you'll learn about past efforts to bring service mesh to Windows, and how Istio's ambient mesh opened the door for a maintainable path forward for a sidecarless deployment topology. We'll learn how Rust, Envoy, and maybe even eBPF can bridge the gap between Linux and Windows for service mesh.

Transitioning from legacy monolithic applications to a modern service mesh comes with its own set of challenges, especially at scale. In this session, Yashwanth and Venkat will take you behind the scenes to explore how eBay adopted Istio for business-critical use cases previously reliant on hardware load balancers. They’ll share hard-earned insights and lessons learned from operating Istio across hundreds of clusters, tackling unique challenges in traffic distribution, and ensuring seamless business continuity during the transition.
The talk will dive into technical details, including the shift from hardware to Istio-powered service mesh, OS and kernel-level optimizations for scaling Istio to meet demanding business needs, and solutions to common pitfalls. This session offers an invaluable opportunity to learn from real-world implementations of Istio and gain practical insights on what works,and what doesn’t, when deploying service mesh at scale.

We will showcase how Kiali 2.0 enhances support for Istio Ambient Mesh and the Gateway API through a demo, providing advanced observability and management for Kubernetes environments. Deploying a sample application with traffic and network policies, we’ll use Kiali's interface to gain insight into traffic paths, Ambient Mesh components, and Istio configurations. With Kiali's new configuration wizards, managing Gateway API routing is streamlined, making it simple to set up and adjust traffic routes. By the end of the session, you'll gain a clear, simplified understanding of mesh infrastructure using Istio Ambient, demystifying the complexity of the Service Mesh.

Considering an Istio multi-cluster setup? Hear real-world lessons on balancing security with operational complexity, scaling configurations to hundreds of clusters, and designing for resilience and observability. Lessons learned from managing 244 clusters serving 570 million requests per day.

Adopting Istio in Kubernetes deployments often starts as an exciting journey, but challenges arise when scaling to support millions of services—issues not covered in basic guides.

This talk dives into the complexities of implementing a multi-cluster Istio service mesh at scale, covering a hub-and-spoke model. Key challenges include: ensuring secure client isolation; load balancing across different types of clusters; managing failover; and automating cluster lifecycle.

Whether you're planning your first multi-cluster Istio deployment or struggling with management at scale, this talk will provide actionable strategies to overcome the pitfalls of large-scale service mesh implementations.

Service Mesh unlocks powerful capabilities for Kubernetes, enhancing observability, security, and traffic control. However, concerns about complexity and overhead have slowed adoption. This session debunks those myths! We’ll showcase how Service Mesh technologies have matured, offering approachable solutions with minimal impact. Istio’s Ambient mode exemplifies this shift, minimizing operational complexity and overhead. Many existing benchmarks raise doubts due to conflicting results. This session tackles this head-on! We’ll run a live, representative benchmark at scale (10s of nodes, thousands of Pods, thousands of RPS throughput) and dissect the metrics with the audience.

As enterprises increasingly embrace generative AI, the necessity for secure, scalable cloud platforms grows. Our session explores the intricacies of traffic management, authentication and authorization within multi-tenant environments and how we addressed these through open-source tools.
Discover how Istio played a pivotal role in managing traffic and ensuring data security, ultimately enabling a secure and efficient AI platform that meets enterprise standards.

So you have Istio questions and you're looking for help? Good news! There's a large Istio community out there with considerable experience you can draw on. Bad news! It's full of very busy people who have jobs and deadlines and places to go and dogs to walk. So how can you get the answers your need? In the this ten minute talk we'll tour the places you can go to engage the community, and talk about the best ways to frame your questions so you can get the information you need and get back to what you were doing.

In an Istio service mesh, all workloads have to be continuously kept up to date with the latest cluster configuration. However, only few or no workloads communicate with every other workload. Keeping superfluous information up to date across the whole cluster creates unnecessary load and increases the cluster convergence time, which results in errors.

One way to optimise this behaviour is by leveraging the Istio Sidecar resource. Workloads specify egress hosts for the services they need to communicate with, updates are then sent only when an egress host IP address changes.

This talk will cover GetYourGuide’s adoption of the Istio Sidecar resource - trade offs of the implementation; rolling out high criticality configuration changes in a live cluster.

We will showcase how we reduced cluster convergence by 10x and eliminated all lag related networking errors by rolling out the Istio Sidecar resource across ~300 services without any incidents and with very little time investment.