[SOLVED] Proxmox Alerts and Notifications Not Working

[void307]

Hi


We've been trying to get our cluster to work with both the inbuilt notifications and SMTP (MS365 and SMTP2GO) config via GUI but it just doesn't generate any alerts when a node or link goes down, user successful/failed login etc.

Any advice besides using 3rd party services and NMS soft like Zabbix/Nagios for now. We just want to get the basics underway first. The SMTP config definitly works as we use it for other services as well.

WE've been up and down on the posted WIKI in below link but no success.

https://pve.proxmox.com/pve-docs/chapter-notifications.html

Last thing needed before the cluster goes into a production environment and getting a license.

 

[cheiss]

Hi,

first of:

user successful/failed login

There is no event for this. Please see the list of available Notifications Events.

when a node or link goes down

There is an event for when Fencing occurs, but not general node/link down if it does not trigger fencing.

Also, have you looked at the system log (journalctl -b) and checked for any email sending failure?

 

[void307]

Hi,

first of:

There is no event for this. Please see the list of available Notifications Events.


There is an event for when Fencing occurs, but not general node/link down if it does not trigger fencing.

Also, have you looked at the system log (journalctl -b) and checked for any email sending failure?

Hi Cheiss

What is a good way to alert for the above? An NMS such as Zabbix etc? This will be one of many clusters that will go in various data centres and we want to ensure everything is in place. Granular alerts for things such as storage, high cpu/memory, nodes offline, link offline, VM's offline etc would be great. I understand from your comment it is not possible within Proxmox itself but rather a 3rd party suggestion?

 

[cheiss]

Zabbix is used quite often to monitor Proxmox (VE) hosts, as far as one can see from the forum. At least it often comes up.
Proxmox VE also features native export of metrics data into Graphite and InfluxDB, as documented in our admin guide: https://localhost:8006/pve-docs/pve-admin-guide.html#external_metric_server

There is also prometheus-pve-exporter, if you'd like to go use a Prometheus/Grafana/Alertmanager stack

In general, if you search here on the forum and on the internet, you should be able to find a lot of information, normally people use these things as basis and build their own full solution to their liking on top.