Use Case for DaemonSet Sharding?

[jcurtis789]

Hello,

I was going through the documentation and wanted to make sure I understood a valid use case for running kube-state-metrics as a DaemonSet. Is the purpose of this feature to provide large clusters (high worker node/pod count) a more efficient way of reporting pod metrics? I run a few large-ish clusters (100+ nodes/20k+ pods) and am looking to run kube-state-metrics as a DaemonSet with the --use-apiserver-cache flag to expose some metrics about pods/containers.

Thanks!

[CatherineF-dev]

Is the purpose of this feature to provide large clusters (high worker node/pod count) a more efficient way of reporting pod metrics?

Yes, one Daemonset kube-state-metrics to collect pod metrics and one Deployment kube-state-metrics to collect other metrics.