Discover How to Monitor Proxmox Cluster!

[@TheMiguel1024]

Thank you, was able to follow along

[@pierre-c9t]

If you run into the API Service not available error, delete your host and instead of adding an agent add an additional macro {$PVE.URL.HOST} and point that to a node in your cluster

[@jenswymeersch9653]

Effectively, you need to use the doc instead of adding agents

[@irdixhiani9317]

Hi I followed each of your steps precisely for my cluster of 3 nodes, but the item count is still 3, and its showing a trigger of authorization issue. Do you have any advice.

[@MRPtech]

On which page do you see item count = 3 ?

[@Josh-mo2ib]

How do I exclude specific VMs (or have an include-only)? I can't find a clear answer on this and I get alerts that things are wrong that I'd like to ignore (e.g. VMs being off)

[@MRPtech]

I am in a same boat as you. Trying to figure this out. I have some VMs that are OFF most of the time and only comes to life when i need them.

[@Josh-mo2ib]

@@MRPtech Appreciate the reply! I'll reply back if I figure it out :)

[@MRPtech]

Hi, I think i got it. Login to Zabbix. Navigate to "Data Collection" > "Hosts". Click on "Triggers" next to your Proxmox node/cluster. Find a trigger you don't want to use and select it. At the bottom of the list click "Disable" this way trigger gets turned off. This makes trigger non-responsive which (i hope) means that no more notifications about VM/LXC being offline

[@Josh-mo2ib]

@@MRPtech Awesome, that worked - thankyou so much!

[@sking379]

Can i express how much i love your channel, techy channels should be like yours, straight to the point properly explained and detailed, lots of techy channels skip through a lot of the setup process thereby causing lots of confusion, although we hsould all do our own research, I totally understand that!!! Keep doing what you are doing, we sincerely appreciate your work Mr. MRP

[@MikeDeVincentis]

Very nice. I used your other Zabbix video to install and configure Zabbix and have set up a few hosts within there and even pushed out a few agents to some systems. This will be my next tackle, to monitor the cluster. Question, is there any benefit of individually adding a vm or lxc using the agent, or does monitoring the host / cluster give you the same results?

[@MRPtech]

Monitoring VM/LXC from cluster i get CPU/RAM/Disk/Network stats. If i would monitor each VM separately i would get more data like Disk wait time, System Load, ect. Monitoring all VM/LXC from cluster is enought data to see what is happening, but for stuff like Docker VM i would get Zabbix agent inside to see not only docker VM performance but each docker container stuff too.

[@MikeDeVincentis]

@@MRPtech That's what I was thinking. For hardware metrics, monitor the cluster. For more specifics on the VM/container, install the agent as needed. Makes perfect sense. Thanks for the reply.

[@MRPtech]

Your welcome. I follow this approach: a) do i need to know IF VM/LXC up or down - Cluster monitoring is fine b) do i need to know HOW VM/LXC behaves - install zabbix agent.

[@Miguelxprimo]

Zabbix agent has to be installed on every host made in proxmox (vms or lxcs) and not on the proxmox console itself, right?

[@MRPtech]

You don't need to install agent on each node. Zabbix will pull data via http connection using created user and it's access API token.

[@esmailbosspro8066]

Keep going your the best

[@kristof9497]

Thanks.

[@VAGELISM22678]

Getting API service not available ... done every step you did...

[@MRPtech]

Double check if API is created on Proxmox side and correct API token key is used

[@VAGELISM22678]

Just installed the agent on every host... And works fine.

[@don_damiano2634]

Such a configuration has no right to work. By configuring the agent, you direct traffic to the specified IP address and port 10050, which Proxmox does not listen on. Are you sure you didn't change anything more than what was shown in the video? Did you edit the template? Maybe you had Zabbix Agents installed on those nodes earlier? Try doing the same thing as in the video but on a fresh installation. You will see that it doesn't work.

[@MRPtech]

Full doc from Zabbix: www.zabbix.com/integrations/proxmox Nowhere it is mentioned that you need to install agent on any nodes. ""This template is designed for the effortless deployment of Proxmox VE monitoring by Zabbix via HTTP and doesn't require any external scripts.""

[@don_damiano2634]

@@MRPtech Correct. By doing everything according to the documentation you linked to, everything works. However, in this case you cannot add more than one Proxmox node. You can only add one macro with the host's IP address. My earlier comment concerned the configuration done as you did in the video, i.e. instead of using a macro, you added interfaces (agents) with the IP addresses of the nodes - and it cannot work. That's why I asked if you did something that you didn't show in the video or if you actually installed zabbix agents on all Proxmox nodes.

[@MRPtech]

I did used macros to setup monitoring (time stamp 05:47) Why adding IP? During my testing of this setup, zabbix complained about not having any IP address in the agent list. To add IP address is optional now starting from Zabbix 7.0 (Zabbix 7.0 runs on Main Galaxy Cluster) Demo environments are virtualized inside Galaxy cluster and before each video i reset their state to "fresh" install. Everything i demonstrate happens on "fresh" OS installs.

[@swubutu]

Thanks again MR.P for your Guidance regarding networking, keep doing what you doing 👊

[@robbuurman1667]

Thanks for this great video. It helps me out a lot.

[@gerry2345]

I like this vid... It helps..

[@MarkConstable]

Excellent information rich video, thanks. I'm not entirely convinced the extra resource usage is worth it over relying on the default graphs. I should try it and see if the alerts alone are worth it.

[@MRPtech]

For most cases what Proxmox already showing is good enough for monitoring purpose. Where Zabbix helps me a lot is go back in history of monitoring and try to find out for example why suddenly Docker VM started to use 90%+ RAM or CPU. Out of nowhere Docker VM started to go full tilt with resources usage. Zabbix helped me to go back in monitoring history and pin a time and date when usage went up and i managed to narrow down to a docker container which went nut with internal updates. I am backing up Cluster to Synology NAS and monitoring network traffic i fine tuned when backups should happen to free network traffic for other things.

[@bast7486]

i just get "Proxmox: API service not available" tried multiple times to remove and add the uuser and changing the tokens and so on. Nothing works for me. :(

[@bast7486]

as soon as i wrote this. it all of a sudden started to work..... took about 40 mins of running then all started. Just strange.

[@emisvaldo]

@@bast7486 Verifique se as devidas permissoes foram aplicadas tanto no usuario quanto na API Token

[@aslfjdasl]

Lovely video, thank you! It is easy to follow. However, I encounter the problem that it says "API service not available". Is there a hidden option to turn off/on the API service inside Proxmox? Happy to get educated on this matter.

[@MRPtech]

Check if Proxmox API setup correctly.

[@nadamajdoub1955]

@@MRPtech I have the same problem but everything is already set up correctly. I'm using Zabbix 6.0.33 and Proxmox 8.2.2