VMware Cloud Community
sjmh
Enthusiast
Enthusiast

Alerting for groups

I didn't see anything in the enhancement requests in JIRA, thought I'd throw up a quick note on the forum here.

Are there any plans for better alerting on compatible groups in 4.0? Right now, you don't get the extra alert criteria, like alerting when it reaches the mark for x min within y min - which means group alerting isn't viable for CPU/memory/load type alerts.

Thanks.
0 Kudos
5 Replies
ama_hyperic
Hot Shot
Hot Shot

You can get the same functionality with a escalation.

Define the first stage of escalation as 'suppress' and then wait x minutes till the next stage which then notifies someone. This gets you similar functionality in a less obvious way. The notification will only go out on the 2nd stage of the escalation if the alert condition still evaluates to true after that time period.
0 Kudos
sjmh
Enthusiast
Enthusiast

I'm not sure how that would get me the same functionality.

The alert based one is saying 'If within 5 minutes, this occurs for a span of 1 minute, alert'.

The escalation method is saying 'If this alert happens, escalate and wait 5 minutes, and if when the 2nd escalation fires, the case is still true, alert'.

The second one could fire if I was at peak for 1 second at both times the alert checked. The first one is actually checking that the alert stayed for one minute total over a certain time period.

It's not quite the same.
0 Kudos
admin
Immortal
Immortal

Another side-effect of Alex' proposal is that it would still generate
an alert (which would clutter your alert center, logs, etc.)

-- Jon


On May 9, 2008, at 9:58 AM, Steven Hajducko wrote:

> I'm not sure how that would get me the same functionality.
>
> The alert based one is saying 'If within 5 minutes, this occurs for
> a span of 1 minute, alert'.
>
> The escalation method is saying 'If this alert happens, escalate
> and wait 5 minutes, and if when the 2nd escalation fires, the case
> is still true, alert'.
>
> The second one could fire if I was at peak for 1 second at both
> times the alert checked. The first one is actually checking that
> the alert stayed for one minute total over a certain time period.
>
> It's not quite the same.


0 Kudos
nibb
Contributor
Contributor

I dont see any real use for the escalation.
Basically it lets you choose a wait time and then what? To notify someone again. Isnt an escalation suppose to be able to do something besides of just send en email or SMS?. ç

In my installation those are the only actions i can choose do besides put it on wait again.

I would also like to know how to do an alert for groups. For example for all disk metrics on all linux servers would be an example. I just see i have to build an alert for each monitoring metric. How annoying.
0 Kudos
admin
Immortal
Immortal

The ability to create an alert in your example would be through the
Group Alerts feature (http://support.hyperic.com/confluence/display/DOC/Group+Alerts
) available in Hyperic HQ Enterprise.

Charles


0 Kudos