Monitor Azure VM scaleset

[riccardopomato]

Hi guys,

Do someone of you how to monitor Azure VMSS with nagios?
How to manage when new VMs are created or deleted automatically?
The names of vm can changes frequently.
Does Nagios has a solution that fits with VMSS?
Thanks
Riccardo

[cdienger]

Hi riccardopomato,

Looks like the Nagios Log Server boards may be the wrong place for this thread. Are you working with Nagios core or Nagios XI?

A couple of Azure plugins exist at https://exchange.nagios.org/index.php?o ... word=azure which may help with what you're trying to achieve, but I don't have an Azure environment to test with to confirm.

[riccardopomato]

Thanks for your reply.
I'm helping a customer to move his frontend Azure vms to Azure vm scale-set.
At moment I need just understand if It's possible still using Nagios and not a different solution.
I would like to understand if Nagios can manage dynamic "born and death" of vms that is typical behavior of azure vm scaleset.

Thanks in advance.
Riccardo

[scottwilkerson]

You would likely want to use the API as part of the provisioning process when creating items.

This would be similar no matter which platform you use

https://assets.nagios.com/downloads/nag ... gement.pdf

[tacolover101]

Thanks for your reply.
I'm helping a customer to move his frontend Azure vms to Azure vm scale-set.
At moment I need just understand if It's possible still using Nagios and not a different solution.
I would like to understand if Nagios can manage dynamic "born and death" of vms that is typical behavior of azure vm scaleset.

Thanks in advance.
Riccardo

You would likely want to use the API as part of the provisioning process when creating items.

This would be similar no matter which platform you use

https://assets.nagios.com/downloads/nag ... gement.pdf

to add a real world use case to this, you could use Nagios log server for this. here's how:
- poll the azure API for metrics every x interval
- parse out those metric variables into their own identifiable fields
- when x metric is reached, have nagios log server reach out to the API to provision a new VM for your cluster. apply same concept to remove cluster node.
- you will need to build middleware to spin up / down and maintain an array of your current vm / variables, but that's the fun part
- sit back and eat popcorn

[scottwilkerson]

to add a real world use case to this, you could use Nagios log server for this. here's how:
- poll the azure API for metrics every x interval
- parse out those metric variables into their own identifiable fields
- when x metric is reached, have nagios log server reach out to the API to provision a new VM for your cluster. apply same concept to remove cluster node.
- you will need to build middleware to spin up / down and maintain an array of your current vm / variables, but that's the fun part
- sit back and eat popcorn

Nice example! Thanks!