Is there a way that I can create a monitoring set that only alarms when multiple performance counter thresholds are exceeded at the same time?
Also, is there a way that the total value of two performance counters can be calculated and this used as the threshold for an alarm?
The reason I ask is that for nearly every performance counter I've seen, a high (or low) value by itself isn't enough to tell the full story, meaning it will either trigger too often or too little to be meaningful or useful.
For example, I've seen advice that 'Free & Zero Page List Bytes' is a useful measure of memory pressure. However I've found that in many cases this gets low but isn't indicative of a problem by itself. Further reading (like this interesting article - http://www.windows7update.com/support-files/memorysizingguidancewin7.pdf) reveals that it is only when the SUM of 'Free & Zero Page List Bytes' and 'Standby Cache Reserve Bytes' is under a certain threshold AND the 'pages input / second' is over a certain threshold that it can be established that a system very probably needs more RAM (i.e. its out of memory and its paging to disk at the same time).
There's no way to create a derived counter within Kaseya at the moment, but it would be a really useful feature for me too. I'm currently interested in tracking TCP retransmits for a client - from the Windows performance counters I can get TCP retransmits/sec, and TCP segments/sec, but what I /really/ want to know is the percentage of retransmits, so I can alarm if an endpoint is experiencing an issue, e.g. "% Segments Retransmitted/sec" = (("Segments Retransmitted/sec" / "Segments/sec") * 100)
I logged a feature request, and it's currently "being considered for a future release", but the more people who ask for it the better ;)