This should improve performance if timeseries starts or ends on the selected time range
vm_concurrent_
vm_concurrent_addrows_*