[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] Monitoring Xen via Nagios


  • To: Xen-users <xen-users@xxxxxxxxxxxxx>
  • From: Simon Hobson <linux@xxxxxxxxxxxxxxxx>
  • Date: Sat, 16 Jul 2016 20:39:57 +0100
  • Delivery-date: Sat, 16 Jul 2016 19:40:57 +0000
  • List-id: Xen user discussion <xen-users.lists.xen.org>

JP Pozzi <jpp@xxxxxxxxxxxxxxxxxx> wrote:

> You will have to learn things about Nagios :
> - Nagios has to know about the hosts you want to monitor
> - Nagios has to know what services yuo will monitor on every host
> So you will have to describe all your hosts and services plus the limits
> (warning or critical) for everything.

Beat me to it ...

To give the OP a bit of a starter ...
Typically you will designate one system to do your monitoring - this does not 
have to be one of your Xen hosts. In my case I run a dedicated machine for 
this, but you could run it as a Xen guest with the obvious limitation that if 
the host goes down then so does the monitoring.

The base install of Nagios will probably only be set to monitor the local 
machine (CPU load, RAM usage, logged in users). You need to define anything you 
want monitoring - and it's a really good idea to see what templates can do for 
you. Don't worry too much at the beginning - just experiment a bit, and be 
prepared to undo stuff and do it a different way as you get the hang of it.

To monitor anything but the local machine, means doing stuff across the 
network. A simple "is this host there" can be done with ping (as long as ping 
responses aren't turned off or firewalled). For stuff on the local network, 
it's also possible to monitor ARP mappings (IIRC I had to write my own plugin 
for this) - not really useful for most networks, but I do it on our hosting 
setup so I can detect various things, not least, when someone sets up a device 
with a duplicate IP address.

You can also monitor services (eg http and smtp) to detect if a remote service 
is down.

Beyond that, you need to "do stuff" remotely - and there are two main ways to 
do that.

One is to use SNMP - basically, if the device does SNMP and you know the right 
OID to use, you can monitor anything covered by the SNMP agent on the device. 
That could be detecting a link down on a switch, or the operating mode of a 
UPS, or ...

For "computers", you can also install NRPE - Nagios Remote Plugin Execute. 
Basically a remote tool which can run Nagios plugins on the target device and 
return the results. Typically (the safe way) you pre-define everything on the 
remote device, but it's possible (at the cost of exposing a potential security 
hole) to pass parameters to the NRPE service - allowing arbitrary plugin 
execution (the security hole being the possibility of arbitrary remote code 
execution if you don't secure the communications channel).

So I would start with the basics.
Get Nagios running and check that the builtin checks are working.
Add some hosts (your Xen hosts, and some guests), and define some services - 
ping is a basic one, after that SMTP for mail servers, HTTP for web servers, 
and so on. It's possible to do a lot of monitoring of your Xen system simply by 
monitoring the guests as though they were standalone machines - if they are 
there on the network, then they are running !
If you wanted more detailed monitoring, then you'd be looking at setting up 
NRPE or SNMP on the Xen hosts and/or Guests. Eg, if you want to monitor disk 
space on a guest, use SNMP or NRPE on the guest to do that.

Just don't try and do everything at once. Get some basics working, and go from 
there.


_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxx
https://lists.xen.org/xen-users

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.