IHAC running lot of NFS services to external client, where many of the IO belongs to nfs_other from volume point of view.
THey are using OM to monitor the system.
They are asking which fields at aggr. level (with OM) to monitor so that they can have a historical view of how many IO in total require disk access. It is because if they view from volume level the total IO involve nfs_read, nfs_write, nfs_other... if they add this up it doesn't equal to the total I/O involve disk access (their understanding is nfs_read or nfs_other may actually resolved in cache which doesn't request disk access).