Correct. Disk I/O information simply isn't available via SNMP. As such it's only available as an option for WMI managed nodes. There really are no implications for changing 50+ nodes beyond the loss of historical data for interfaces and volumes. This is because there's no reliable method of mapping SNMP volume and interface index IDs to WMI equivalents. As for performance, there is a slight increase in polling overhead associated with WMI vs SNMP, but with 50 nodes it would be so negligible as to be irrelevant. This increase is normally not noticeable until polling several hundreds or thousands of nodes via WMI.
↧