In small clusters, we should expect a failover to happen right after `down-after-milliseconds`. CI currently indicates it takes longer than that.
In small clusters, we should expect a failover to happen right after
down-after-milliseconds. CI currently indicates it takes longer than that.