always

when viewing dashboard panels within observe -> dashboards 
the dashboards that contain a value for "memory usage" and that query be against `container_memory_working_set_bytes`
the query also contains - "container!=''" 

if you navigate to workloads -> pods -> click pod name -> metrics -> memory usage the query does not contain - container!='' 

it contains container="" 

should these values not match? 

I understand the difference to be: 

sum(container_memory_working_set_bytes{pod=~'POD_NAME',container!=''})  
would usually be expected to be  smaller than container_memory_working_set_bytes{pod=~'POD_NAME',container=''}.  

The former sums up the memory usage of all the managed containers in the pod, the latter reprensents the memory usage of the parent cgroup of the pod, that includes additional resources to manage the pod.