Lightning Talk
summary
Kubernetes replaces per-second polling with a runtime-driven stream so kubelet stops listing containers every second and instead receives gRPC streaming events from CRI-compatible runtimes like containerd cri o or Docker shim. Implementation uses Go channels, a stateful cache, and four event types (created started stopped exited). Result is lower CPU and memory on dense nodes, relist interval rises from ~1s to ~5m, and short-lived Pods now surface a Running state when feature gate: EventedProbe is enabled (beta in 1.27).
AI-written summary · about
related talks
06mLightning Talk: OpenKruise: Advancing Kubernetes Workload Auto...
07mLightning Talk - Say what? You're Running the Storage Platform IN Kubernetes?
11mLightning Talk: How Jenga & Tetris Can Help Run More Containers with Lesser Costs & Downtime
05mProject Lightning Talk: Transforming your Application Behavior to Kubernetes Objects
06mLightning Talk: Beyond Operators: Reimagine Distributed Applications on Kubernetes
06m⚡ Lightning Talk: Colocating Data Workloads and Web Services on Kubernetes to Improve Resource Utilization
05mLightning Talk: Slow Starting Containers, How to Check Their Health?
35mDeep Dive: Kubernetes (Instrumentation) SIG