I'm new to Hyperic and vCops. We have a mostly virtualized environment but I've been installing Hyperic on virtual machine and physical machines in our environment. For virtual machines I want to monitor specific windows services and processes that aren't captured with vCops alone. So I install the Hyperic agent on some machines and setup monitoring of specific windows services. Then I setup the HypericApiAdapter Win32 attribute package to generate a critical alarm when services are not available. I assign this attribute package to all Win32 resource kinds. After that I configured the emailFitler.xml filter to send an email when the Win32 resource kind has a critical alarm. Everything worked appropriately until a network connection to a site dropped. What happens is that I get an alarm that the machine is down, then I get an alarm for the services that I configured for monitoring are down, and then I get additional alarms for every processor, disk drive, and network adapter in the machine being down. Hyperic considers a CPU, network adapter, disk drive, etc as a service in addition to a monitor for a specific windows service. To me that is excessive. If the machine is down why does there need to be notifications for every single service. It almost makes it impossible to use the email functionality from vCops when you are integrating with Hyperic. Has anybody experienced this and how are you working around it? Do you just use the email functionality from Hyperic for Hyperic resources? I wanted to manage everything from vCops but I'm beginning to think this may not be possible?
Thanks for the help.
We configured it on Hyperic and brought the resources into vcops only for dashboarding and not alreting. The alert configuration on Hyperic is better. We have configured all our servers to bring up the services if they go down using the control actions in the hyperic alerting.
My problem with Hyperic is the email functionality is even weaker than in vCops. I have 3 options to choose from.
1) Alarm each time conditions are met, which means an email every time the condition is met. If we have the Win32 platform type set at it's default of 1 minute for the collection interval that means an alarm every minute. I can change to maybe 5 minutes the but an alarm every 5 minutes is excessive.
2) generate one alert and then disable alert definition until fixed. I want more than one alert so this will not work
3) Once every _ times conditions are met within a time period of _ minutes. I don't understand this one even after looking at the documentation. Below is what it says. So if I set the collection interval to 2 minutes for availability on the Win32 platform type and the condition is 3 recurrences in 5 minutes the condition will never be met. The collection interval is 2 minutes. It will only do a collection 2 times in 5 minutes so how would something like this every work. I obviously must not be understanding this properly.