Maximizing performance requires deep node-level customization and precise benchmarking. In this session, explore the configuration and tooling for optimizing stateless, stateful, and inference workloads on Google Kubernetes Engine (GKE). Discover new GKE APIs that generate pod manifests with built-in, industry-standard optimizations. Dive deep into benchmarking and tuning workloads like Nginx, Redis, MySQL to get a preview of packaged optimizations. Learn to use open source frameworks to reproduce these gains for your unique workloads. And leave with a performance-tuning playbook and a reference architecture for high-scale GKE nodes.