ESXi Host State StandBy

thorian93 · July 23, 2020, 6:12am

Hello there!

I got a quite simple request I assume.

Problem

I want to able to map the “StandBy” state of an ESXi Host to a state in checkmk other than UNKN. Obviously this is not possible, although other states can be mapped. This is not a huge problem for me but I think the implementation would be positively recognized by all vSphere admins and operators as “StandBy” is not a UNKN state really.

Feature Description

The ruleset State of ESX hosts and virtual machines should be able to map a “StandBy” state of an ESXi host to another state than UNKN. Preferably this would be freely configurable, as there are use cases, where a “StandBy” state is a WARN or an OK state, or maybe even a CRIT state.

I found these posts in the German archives, just for reference:

Thanks in advance and have a good one!

ttr · July 23, 2020, 1:31pm

Hi @thorian93,

I made some changes to original checks in esx_vsphere_hostsystem to support this. You may want to give it a try, but I won’t support/fix/whatever this. Here it works very well (cmk 1.6, vSphere 6.7).

https://github.com/tretbar/check_mk/blob/master/esx_vsphere_hostsystem_dpm/esx_vsphere_hostsystem_dpm-2.0.mkp

Some doc:

github.com

tretbar/check_mk/blob/master/esx_vsphere_hostsystem_dpm/README.md

## esx_vsphere_hostsystem_dpm

This package extends esx_vsphere_hostsystem.cpu_usage, esx_vsphere_hostsystem.mem_usage, esx_vsphere_hostsystem.state from esx_vsphere_hostsystem to support DPM.
The check script overrides static "esx_vsphere_hostsystem" with script in local file hierarchy, but imports and executes original one to be update-save. 

You want to:
* query both ESXi hosts and vCenter Server
* get host power state from vCenter
* set "host check command" to "Overall state" for DPM hosts
* disable services "HostSystem" and "Disk IO SUMMARY" for DPM hosts

A host in DPM stays OK, CPU and Memory will show an appropriate message.

Regards,
Tom.

mace · July 24, 2020, 5:50am

Hi,
we trigger a downtime with a vcenter action when the host is put into standby or maintainance mode and remove those when it comes back.

the downtime wrapper script is here:

gist.github.com

https://gist.github.com/m4c3/e3d8e3730e7a5575e84008a66c472bfa

action-downtime.sh

#!/bin/bash
# this script creates downtimes in checkmk based on vmware vcenter alarm actions
# put this script in /var/local/ on your vcenter appliance
#
# requires the python3 compatible downtime handler from 
# https://raw.githubusercontent.com/tribe29/checkmk/master/doc/treasures/downtime.python3.py
# in /var/local/

USER="automation"
PASS="secret"

This file has been truncated. show original

the alarm rule looks like this

but i really like the approach of @ttr! i have to try this out
ciao matthias

thorian93 · July 24, 2020, 6:15am

Hi @ttr and thanks for that, looks like a quite nice solution. I would love to see it in the base functionality of checkmk but I assume other issues are more pressing.

I will give your solution a try, thanks a lot!

thorian93 · July 24, 2020, 6:18am

Hi @mace and thanks for your answer. I see the approach but I would rather solve this from checkmk without interacting with the vSphere admins. Not that I don’t like them, I myself was one, but the solution of @ttr looks more scalable and easier to handle for me.