I have an alert that is set up as shown
Which I plan to have email out certain details as shown here.
When I test email notifications I get this
Which is not quite what I expect. Any help would be appreciated
When you are sending your test email alert, are you selecting a Volume in the Test Fire Alert window as well as the node? If you are only selecting a Network Node and do the test, then you get a reply like you did.
I see you have a condition for the "percent used to be equal to 95.01919". Seems like it would be tough to hit this on the button. Try change "is equal to" to "greater or equal" and give it a try. Then the variables may be able to match the volume in question.
Zak Kahl
Loop1 Systems
http://www.loop1systems.com
You would think the 95% mark would be hard to hit, but the app logging goes to /opt (yes, I know that is a poor design choice, no I do not have any say in the matter) and fills rather quickly. Also, the log rotation is not, how can I put this, well thought out.
I will try lowering the threshold and test again. I will post results when I am back at my desk.
Thanks
No joy,
set the condition to >= 0 and tested against an active server
One more thing you could try. In email actions add "Volume." to each variable. This will point it at the volume table.
${Volume.VolumePercentUsed}
Same result, I am at a loss as to where the disconnect is.
Lets try removing the condition for the volume descriptions. Just so we can see those variables getting filled in.
This looks more of an issue of the testing than the alert message.
What are you putting for your test node/interface/volume?
Remember that it takes the variables from the table of what you specify. So if you have InterfaceID but are alerting from a node, it'll just display ${InterfaceID}. Also, remember that custom variables are specificed as being either node or interface properties. Not both. Look to see if you can edit it with "edit node" or "edit interface" option.
Lastily, remember that testing ignore conditions and restrictions and goes straight to actions. This means that you can't really test to see if your alert conditions with test alerts, only the information when it sends is correct (so it won't catch that =95.0... makes your alert very unlikely to fire).
That was it, I assumed the test would scan the node and report on any of the disk volumes. Now I know better, thanks you
this seems to be resolved, a quick question regarding alerts. I have multiple alerts that I have configured for various potential issues. (I know this is a surprise to everyone) I test an APM: Hardware monitor alert on a box that I have offlined a drive on, the test doesn't seem to pick up this down drive is that normal for testing alerts?
Test alerts are just that, it sends a generic test alert. It doesn't actually check if there's anything wrong with the device.
One other question, on alert suppression I would like to not be notified when our DEV servers are full or approaching full. I have all of our dev servers in a group named DEV/QA, I would like to exclude this group from alerting but can't seem to find a way to use the group is equal to DEV/QA in the suppression options. When I try to change the trigger condition so I can use the groups I get the error "Please specify a complete Trigger Condition Field not specified or invalid" I could go in and add the individual nodes but that defeats the purpose (at least as far as alerting is concerned) of grouping servers together.
Is there a way that I can get this to work using the Group field?
If you have ready any of my posts, you know I'm a custom property junky.
I've got a node field called "Prod_State" that includes Prod (duh), Dev, QA, Test, etc.
My groups use THAT field to determine what's in it (that way my lower level techs don't need muck around with groups. They can just change the custom property to make something go where I want it).
So my disk alert includes a check for "if prod_state = 'PROD' " and that eliminates unwanted wake-up calls for dev boxes.
Not QUITE what you were asking, but it'll get the job done.
It looks like this is the only way to accomplish that. Thanks
