ESX Heartbeat warning


I’m monitoring my esxi 6.7 host and all vm’s on it with CheckMK. From time to time I am getting an Heartbeat warning which disapears one minute later. I doesn’t seem as if the server is under pressure at this time.

How to find out, what is going wrong there and how to fix the issue?

Any help would be greatly appreciated.

You vCenter means that it cannot communicate with the virtual machine.
Is it possible that this happens every time you do some backup jobs inside your ESX environment?

No, it happens from time to time and on different machines (Linux and Windows). But I am not able to see a concrete scheme.

Do you query the ESX host directly or a vCenter machine?

I do both like I found in a tutorial of CheckMK.

Anyone a idea of this? The bahaviour is reply annoying.

Like i said before this is a problem inside your ESX environment. If you use CMK 1.6 it is enough to query the vCenter, the ESX host don’t need to be queried.
Since 1.6 you also get the performance counters for the hosts from the vCenter.
If the Heartbeat warning occurs again then you should have a look at your environment setup.

I’m going to disagree with this. As Vcenter is not necessary for VMs. Vcenter can be downed and rebooted, etc. The piggyback through the ESXi is likely more important. A VM has be on an active hypervisor (ESXi) node. Vcenter is just an add on orchestrator.

? If your vCenter is not working all VMs will run yes. But no HA failover or no backup is possible. Your vCenter is the central part of your environment. One other point you should remember is that you don’t need on every host a own user for the monitoring. I have ESX environment with over 100 hosts and no one from the admin team wanted to create the local users :slight_smile:

It’s designed to be down. Because the implementation is an “after the fact” (sort of a hack to be honest) implementation, a VM “running” doesn’t depend on it at all. The key is actually the hypervisor itself. Vcenter is merely an abstract out of band orchestrator that attempts to move things around on top of a pseudo-cluster. So, I stand behind the idea that ESXi is the critical piece where the VM runs and Vcenter, while it is very very interesting, it’s not the critical dependent piece.

Now, if what is being monitored is the capability of moving hosts around (for example), certainly, that is where Vcenter comes in. Or anything else “orchestrator” related.

Exaggeration ahead, it’s like saying a programmer can’t program without a high resolution twin set of monitors. Some might say, “that’s true”.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.