Solved

Alert on VM Performance

  • 10 May 2016
  • 7 replies
  • 11942 views

Badge +3
I recently had an individual VM whose CPU spiked to between 94-98% for roughly 6 hours. I'd only caught it because the user finally called in and said their VM was slow. In Prism it wasn't showing a warning at all, but I could very clearly see that the CPU was pegged.

I wondered if there was any way to set up an email alert for if CPU/RAM at the VM-level has sustained utilization above a certain threshold?
icon

Best answer by harryhy 11 May 2016, 20:37

Right now, the cluster health has a check for the user VM cpu and memory. We will add threshold based alert for user vm metrics in the future releases.
View original

This topic has been closed for comments

7 replies

Userlevel 7
Badge +30
Was this for an application hosted on Community Edition? or the commercial version of our software?
Badge +3
Ah, maybe I'm not in the right forum!

This is for the enterprise version of Prism - we have a 3-node Nutanix cluster hosting about 70 virtual desktops.
Userlevel 7
Badge +30
ok, I moved this to the right forum (Prism).

For customized alerting and so on, you'd be looking for Prism Pro, which is built into Prism Central. You can deploy this and check it out even on a single cluster system.

Prism Pro is caulk full of all sorts of cool stuff, not just neat alerts, but capacity management, trending, tagging and so on.

Give it a look and see if it suits what you're thinking about.


The other way to do this is, from a very basic A|B perspective, is set up a CPU usage alarm in vCenter (assuming you're using VMW) and have it email you when it hits certain levels.
Userlevel 2
Badge +11
Right now, the cluster health has a check for the user VM cpu and memory. We will add threshold based alert for user vm metrics in the future releases.
Userlevel 3
Badge +17
Hi Jon、

Currently my customer is also asking the same question as they want to use Prism to have the email alert sent back to them each time the CVM or VM CPU, RAM or disk usage running exceed a specified number of percentage of performance

I have also had a look around Alert Policies but there is no entry that related to those kinds of VM performance as I wrote above as well as in Prism Central

Do you have any idea on this?

I think it would be a good idea if Nutanix could add more policies to Alert field for development and to suite with the customer request in the future
Userlevel 7
Badge +30
Sure, I see what you're saying. We're adding new features all the time. I'd encourage you to check out AOS 5.0 and the associated Prism Central when it is released (due out soon), as we've released a TON of new features in there.

I dont believe this specific one is in there, but check it out either way, and you may find things close to what you're looking for.
So were threshold based alerts on VM performance ever added to any AOS versions by any chance? Can Prism Pro do this?

We would like to receive email alerts on High CPU/RAM from our AHV cluster like we do from our ESXi cluster through vCenter.