Do Not Disturb Mode - Fine-Tuning Resources for Latency Sensitive Workloads

Explore the intricacies of fine-tuning resources for latency-sensitive workloads in Kubernetes environments through this 21-minute conference talk presented by Antti Kervinen from Intel and Peter Hunt from Red Hat at a CNCF event. Discover how recent extensions to CRI-O and containerd enable administrators to control resource sharing beyond dedicated CPUs, including CPU cache, memory bandwidth, and storage I/O. Learn about the hardware requirements and configuration techniques that allow applications to be categorized based on their resource needs. Gain insights into optimizing workload performance by minimizing interruptions and resource contention, ultimately improving the efficiency of Kubernetes deployments for latency-sensitive applications.