Hi Kenan,
In the query block a, can you please try the uid_in function and check if it improves performance.
Here is an example:
{
devices as var(func: type(Device), first: 1000) @filter(has(has_object)){
h as has_object (first:10)
}
q(func: uid(devices)) @filter(uid_in(has_object,uid(h) )) {
name
count
has_object {
name
count
}
}
}
Please check if adding the has query also improves the overall performance.