Script triggered by alert, also monitoring Xen under Linux

[jantman]

Hi, I've just started testing Hyperic and some other NMS's on my home development network before recommending one to roll out at work. I have 2 questions (I'm using Hyperic HQ, the Open Source version)

1) Is there any way to add calling a script on the Hyperic server as an alert action? For some platforms/services, I'd like to automatically create a ticket in my ticketing system on an alert (if it goes un-acknowledged for an hour). Is there any way to add an action that calls a script and passes it some relevant alert information?

2) I searched HyperFORGE, but can't seem to find any plugins that monitor Xen domains under Linux. Any suggestions? I'd like to get status, process and memory usage, etc. It seems quite an oversight to have a product that's largely Unix-based and only have support for VMware (this is a make-or-break issue for me, as most of our architecture is being migrated to Xen-based machines).

Thanks,
Jason

[BradFelmey]

Your first question is something called a control action. It's a fantastic feature of the Enterprise version. 😉

Poor man's way of doing this is to have an alert sent to a box that can use something like procmail to fire off the desired action.

For your second question, I presume you're talking about something akin to the VMware Infrastructure Client's representation of guest resource utilization?

[jantman]

As to the script - I guess that would work, but it seems like a lot of work for something that's a *really* simple feature. Honestly, I'm a bit confused at the feature difference between Open Source and Enterprise - Open Source has a number of really advanced features, but is missing something as simple as alert scripts, contact groups, and time-based alerting?

As to the Xen monitoring - I don't know anything about VMware - the entire reason why I'm trying Hyperic is that we're an entirely Open Source shop. So I don't have any experience with VMware.

I was talking about something akin to XenMan, or even something that queries the output of "xm" which shows machine state, time up, memory used, etc.

[JohnMarkOrg]

Hi there,

I've heard rumors of a new Xen plugin being available "real soon
now", but I'm not sure what the target release is. I can see about
putting the current snapshot on the HyperFORGE, if you're up for
testing, but first I have to inquire as to the state of said plugin.

-John Mark


Excerpts from Jason Antman's message of Fri Feb 08 10:20:54 -0800 2008:
> As to the script - I guess that would work, but it seems like a lot of work for
> something that's a *really* simple feature. Honestly, I'm a bit confused at the
> feature difference between Open Source and Enterprise - Open Source has a
> number of really advanced features, but is missing something as simple as alert
> scripts, contact groups, and time-based alerting?
>
> As to the Xen monitoring - I don't know anything about VMware - the entire
> reason why I'm trying Hyperic is that we're an entirely Open Source shop. So I
> don't have any experience with VMware.
>
> I was talking about something akin to XenMan, or even something that queries
> the output of "xm" which shows machine state, time up, memory used, etc.

[jantman]

Thanks for the info. I'd be more than happy to give it a test. At the moment, I'm evaluating a number of systems, and currently have the current HypericHQ running, as well as OpenNMS, and will have the beta of the newest GroundWork Open Source. I'll probably give Zabbix a spin, too.

I'm going to write out a bit of a review, but for my purposes, I have a set list of needed features, which include:
1) Monitoring of HP Proliant hardware and SmartArray RAID status
2) Monitoring Xen status
3) Trigger a script on alert to automatically open a ticket for some problems.

[roger_symonds]

Hi jantman,

I was checking your blog (http://jantman.dyndns.org:10011/blog/2008/02/foss-monitoring-comparison-hyperic-part.html) and was wondering if you've had a chance to put up a third installment?

Roger