2026-10-01 –, Main Hall All times in Europe/Vilnius
Vinted’s platform scales dramatically, supporting more than 300,000 external requests per second to core service and executing over 5,000 deployments daily in all Kubernetes clusters.
This session dives into how Vinted successfully built and operates this infrastructure, which includes 20 bare metal K8s clusters spanning 2,500+ physical nodes and 310,000 CPU cores.
I'll share lessons learned across four core pillars:
- Scalability: Operating bare metal K8s clusters at massive scale
- Reliability: Implementing Istio multi-primary service mesh for traffic shaping, failover handling and sharing critical performance optimizations
- Velocity: Accelerating CI/CD with a GitOps workflow (ArgoCD and Argo Rollouts) that features automated canary deployments and analysis
- Productivity: Enhancing developer experience and self-service capabilities through a Backstage-based DevPortal
I am a passionate technologist with a focus on cloud-native technologies, particularly Kubernetes (CKA, CKAD, CKS, KCNA, KCSA) and Istio. As a co-organizer of the Cloud Native Lithuania meetup and a CNCF ambassador, I am deeply involved in the cloud-native community. Leveraging my expertise, I currently serve as a Lead SRE at Vinted, contributing to the reliability and scalability of our systems, while also sharing my experiences as a speaker at technology conferences.