For windows servers where you cannot install an agent (due to policy), but have SNMP...
Has anyone found a practical answer to memory usage alerting?
Monitoring raw usage is mostly pointless, since windows (appropriately) tries to use it all as it gets busy. "Normal" periods can easily hit 95, 98%, and you get too many false positives.
The key to me is hard page fault rates over time. I can get that in Linux from SNMP, but not windows. At least I cannot find anywhere I can get it.
There's numerous postings from the WNT/W2000 time period about OID's that provide it, but none I've seen posted work. Not sure if they are just wrong, or it changed in later versions. Yes, I know SNMP is officially deprecated in Windows, but sometimes it is what we have to work with.
Am I correct it just isn't in there?
Anyone find a better method for identifying real issues with windows memory usage via SNMP alone?
Has anyone found a practical answer to memory usage alerting?
Monitoring raw usage is mostly pointless, since windows (appropriately) tries to use it all as it gets busy. "Normal" periods can easily hit 95, 98%, and you get too many false positives.
The key to me is hard page fault rates over time. I can get that in Linux from SNMP, but not windows. At least I cannot find anywhere I can get it.
There's numerous postings from the WNT/W2000 time period about OID's that provide it, but none I've seen posted work. Not sure if they are just wrong, or it changed in later versions. Yes, I know SNMP is officially deprecated in Windows, but sometimes it is what we have to work with.
Am I correct it just isn't in there?
Anyone find a better method for identifying real issues with windows memory usage via SNMP alone?