Skip to content
Documentation

Documentation

Drop pre-caches container images on Kubernetes nodes using short-lived Pods.

Why

When many CI jobs or workloads start simultaneously, Kubernetes nodes face a thundering herd of image pulls. Concurrent pods on the same node all pulling the same large image saturate bandwidth, stall containerd, and cascade into failures.

ProblemImpact
Thundering herdParallel pulls of the same image destabilize nodes
Registry overloadSudden pull surges hit rate limits or cause outages
Cold-start latencyLarge images delay workloads that need them immediately

Drop pre-caches images before workloads need them, paces pulls to stay within safe limits, and automatically discovers which images matter most.

Sections

SectionWhat you’ll find
InstallationHelm install, prerequisites
UsageCachedImage, CachedImageSet, PullPolicy examples
DiscoveryAutomatic image discovery with DiscoveryPolicy
MonitoringPrometheus metrics, events, status conditions
ReferenceCRD field reference, errors, metrics, architecture
DevelopingBuild, test, lint, project structure
For AI Agentsllms.txt, Markdown API, generation architecture
Last updated on