Alert threshold overriding

FormerMember

Hi all,

I'm implementing an NPM and APM environment to monitor many hundreds of servers.

I'm trying to work out the best way to alert on servers. Most servers will use the default thresholds for disk space, memory etc. but some machines will either want to ignore those alerts or override the default with a different value (i.e. 99% disk utilized rather than 95%).

For example for some Oracle servers, the memory utilization will always be 100% (preallocated) so we'll want to ignore this in the alert. On another machine a page file is preallocated on a disk to use 98% so we'll want to raise the threshold of the disk utilization alert for this particular server.

If i want to do this do i need to create a new alert for each machine which we need different thresholds for? This list could become huge if this is the case. I could potentially create a complex rule for the default alert which triggers on different thresholds but this would become a massive rule if we add lots of overrides to it (and i assume probably won't perform very well).

Does anyone have an experience on this and can provide some tips/advice?

 

Thanks,

Chris.

jcooler

I think I know how you can accomplish this. You could create a custom property for volumes. Its say you call it "criticalDiskUsage".You then set the percent that you want to get alerted at. For the devices that are always 100% full, set the custom property to 101. for all others, set it to the desired level you want to get alerted at. (99, 95 90 etc.)

Then, create an advanced alert. Under trigger condition, choose volumes as the property to monitor, then add a complex condition that reads something like this:

field   Volume Percent Used    greater or equal   field    criticalDiskUsage

 

I haven't tried this but I think it would work for what you need to do.  Adding the custom property values make take a little bit of time, but its not that bad if you add them using the Custom Property Editor.

 

Hope this helps you out some.