[Verse 1] Inside the cluster, pods awaken slow Pending phase first, then scheduling flow Kubelet pulls images, containers spawn Running state reached when all deps are drawn But failure lurks in restart policy OnFailure, Always, or Never decree [Chorus] Pods hold containers, runtime decides CPU and memory, resource guides Init before main, sidecar beside Kubelet orchestrates the container ride Limits and requests keep workloads aligned Pod lifecycle burns bright in my mind [Verse 2] Container runtime interface calls CRI plugins answer when kubelet crawls Containerd, CRI-O handle the load Pull, create, start follows runtime code Init containers run sequential, clean Main containers launch parallel scene [Chorus] Pods hold containers, runtime decides CPU and memory, resource guides Init before main, sidecar beside Kubelet orchestrates the container ride Limits and requests keep workloads aligned Pod lifecycle burns bright in my mind [Bridge] Quality of service governs the queue Guaranteed pods get resources true Burstable scales when headroom allows Best effort dies when pressure growls Node pressure triggers, eviction strikes Failed pods restart per policy likes [Verse 3] Readiness probe guards the traffic gate Liveness probe kills when containers wait Startup probe gives slow apps more time Termination grace period marks decline SIGTERM then SIGKILL ends the show Pod deletion cascade, watchers know [Chorus] Pods hold containers, runtime decides CPU and memory, resource guides Init before main, sidecar beside Kubelet orchestrates the container ride Limits and requests keep workloads aligned Pod lifecycle burns bright in my mind [Outro] From pending through running to termination Kubelet manages container creation Resource quotas, namespaces contain Pod patterns repeating, the cluster's refrain
← Kubernetes Core Architecture | Kubernetes Workload Management →