vsphere monitoring

[lukesullivan]

I'm using the vmware wizard to add my vcenter server. I'm only interested in the status of the esx hosts themselves, not any of the guest machines.

Vsphere knows about all of the esx hosts connected to it (all esx hosts roll performance data up to vsphere).

however it looks like I have to re-invent the wheel and add every esx host individually to nagios (somewhat tedious, I have 60+ esx hosts).

Is there something I'm missing? Is there a feature that lets you run the wizard against a vsphere server, and use that as the monitor for all the esx host stati? If there is no feature, could I suggest that it would be a good one to have?

thanks,

-Luke

[npolovenko]

Hello, @lukesullivan.Would all 60 ESX hosts work with the same username and password? If so you can run a wizard against one ESX host and then assign its services to multiple hosts using the manage host button. You can clone host configs using the "clone" button in the CCM, and then change the IP address and the host name.

[lmiltchev]

Just to add to what @npolovenko said - you could set up a host the way you want it and use it as a "template", cloning it into many other hosts via the Bulk Host Cloning and Import wizard (provided you are using the same username/password, and you are monitoring the same services on each of the "clone").

Hope this helps.

[lukesullivan]

.... that just seems awkward if there is already a place that rolls up all the performance data. Rather than 60 hosts, I can have one host.

That's kind of the point of vsphere.

[lmiltchev]

Rather than 60 hosts, I can have one host.

Can you elaborate on that? What kind of metrics you would like to monitor on each host? If you want to run the wizard against the vcenter and have one host ONLY, what kind of services you would expect to have on this host?

For each of the 60 hosts that you have, you will have several services, e.g. VM Status, Networking, Services, Datastore Usage, CPU, and Memory... These are quite many checks to have on one host... Am I missing the point here?

[lukesullivan]

yes, that is some number of checks on one host. I've got other hosts that have 100+ checks though (interface status on some of the routers in my environment...). Also, the total number of checks that need to be run are the same (ie: if I have 5 esx hosts, and I want to check cpu and memory on all of them, that's 10 services, it doesnt matter if that's 10 services on one host or two services per host over 5 hosts...)

partly what I'm after is reducing the connectivity. My esx hosts span multiple datacenters, so would need several acls in place to access all of them. If I could just have nagios ask one system, since that one system already has all of the data needed to answer for all of the hosts that I have, I would have a simpler overall structure.

I dont think it's a huge deal, what nagios provides out of the box is probably fine for the majority of customers, and I'm probably an outlier.

You can go ahead and resolve this issue.

thanks,

-Luke

[lmiltchev]

You can go ahead and resolve this issue.

I will be closing this topic then. If you have any further questions, please start a new thread. Thank you!