This came out of the https://bugzilla.redhat.com/show_bug.cgi?id=1943704. 

Add dashboard for iowait CPU on master nodes, this will help customers and customer support or us identify problems that result in leader election - we can see that often due to high iowait, aligning with large spikes in fsync and or peer to peer latency.

Query:

(sum(irate(node_cpu_seconds_total {mode="iowait"} [2m])) without (cpu)) / count(node_cpu_seconds_total) without (cpu) * 100
AND on (instance) label_replace( kube_node_role{role="master"}, "instance", "$1", "node", "(.+)" )