what alert are you trying to create?
currently i monitored 7 cluster and each is on a 7 different location/country. i would like to create alert for:
1. alert which based/categorize on their location/cluster.
if a VM123 is down/poweredOFF, an alert will trigger " a VM123 from HostA or ClusterA is down/poweredOFF" and this alert will then email to Support team A.
the reason is, i need to aasign the support email address based the node location. currently i managed to create/modify an alert but i dont know how to categorize/filter by cluster/host
Let's walk through it
One way to do this is to monitor the hearbeat on the VM that is communicated from VM tools to the host. Note that VM tools must be installed on the guest to properly receive the hearbeat metric.
So we want to know if a VM has no heartbeat OR of the VM is powered off. The query for that is
(vm.heartbeat.latest:0 OR vm.powerstate:poweredOff)
Now we want to create an alert for each cluster, so that the alert can be emailed to the right recipient. Let's walk through setting up the alert for ClusterA. Now the query becomes
(vm.heartbeat.latest:0 OR vm.powerstate:poweredOff) AND vm.cluster.name:ClusterA
This will only return the VMs in cluster A that have no hearbeat or are PoweredOff.
Now we have the right query, we can click on "alert" to create an alert
This will take us to the alert configuration screen. On the 1st tab, I should check evaluate on performance collection, since both Powerstate and Hearbeat are performance metrics so this alert will get computed every time performance data is collected (every 10 mins by default)
The criteria tab is next, and that looks fine since we want this alert to fire if any VMs are returned by our search.
On the notifications tab, you have the option to define the properties that show up when an alert fires - I want the Power State of the VM and the heartbeat value (which is a %) to show up. You can also see that this tab is where the Email address is defined that you want the alert sent to.
Attaching the Alert created in the above example, example_alert.xml, this can be imported into Virtualization Manager in the "Content" tab inside the product.
You will need to edit the cluster name to match your environment, and edit the email address that you want the alert sent to.
You can do a "save as" on this alert to create one for each cluster/site that you need.
Hi there Jon,
Thank you soo much on the step by step guide....it really helps especially on the use of heartbeat / vm tools...i got it applied on all of my cluster..thanks!
the only problem is the 'performance' data interval is 10 minutes... meaning i will get the alert after 10 mins/or the next poll...any impact if i set the interval to 2 min or lesser?
anyway that i can get a refference for the alert creation? what language it use?
There is a section on Alerts in the online manual here http://www.solarwinds.com/documentation/vMan/docs/html/en-US/wwhelp/wwhimpl/js/html/wwhelp.htm
You can lower the collection schedule for performance jobs to a minimum of 5 minutes
You probably want to make sure that your performance jobs are taking much less than 5 mins to complete in your environment before doing that
thanks, all cleared!..
now only when the vm rebooted...vmanager cant even detected and triggered alert..any ideas for this?
thanks a lot!