Proxmox monitoring / alerting question

[tlex]

One of the reason why I still use TrueNas to manage my disks with a passthrough in Proxmox is because if there is a disk failure or anything, I can easily get an alert by email.
I was wondering if there is any alternative directly in Proxmox in order to customize or at least define what kind of event / alert could trigger an email directly from Proxmox ? Or how do you guys deal with it ? Basically what I'm looking for is a bit like the "Alerts" in Esx...

 

[RokaKen]

You mean something like https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_disk_health_monitoring ?

 

[tlex]

You mean something like https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_disk_health_monitoring ?

Correct. But I really hope we could have a bit more control on the types of alerts, including global system status, temperature, cpus, vm reboots, etc... I'm just surprised that is not implemented out of the box like it is with Esx. Is this something Proxmox is working on ?

 

[ness1602]

This is what classic NMS is used for, or newer something like Influx and Grafana.

 

[tlex]

This is what classic NMS is used for, or newer something like Influx and Grafana.

Nah... I don't think so... these are for performance mon aren't they ? Have you ever seen Esx alerts ?

 

[ness1602]

Everthing is performance, even status,temp ,cpus etc.

 

[birdy]

There are many ways to get a job done, but if my main goal was monitoring/alerting I probably would not use Grafana/Influx because they are mainly targeted at gathering and analysis of lots of data. Proper monitoring tools are for example Zabbix or PRTG (and many more). But as I think you are looking for a rather simple/basic solution you could look into this one:

https://www.netdata.cloud/

 

[gurubert]

IMHO monitoring and alerting is not the job of a virtualization system like Proxmox.

Use an infrastructure monitoring system for this job.

 

[Dunuin]

I use Zabbix for that. Its running inside a VM with 3GB RAM, 2 vcores and is writing about 70GB per day to the disk to monitor about 30 VMs/hosts. There are alot of pugins and you can even use it to monitor your TrueNAS servers using SNMP. A little bit annoying is that you can't monitor the SMART data of the TrueNAS hosts disks because TrueNAS is a appliance and you aren't allow to install stuff. So the zabbix agent can't be installed to pull data using smartctl directly and SNMP isn't reporting SMART data. I've read about some people trying to run the zabbix agent inside a jail but they didn't got it running for SMART because the jail had no access to the hosts drives.

 

[birdy]

IMHO monitoring and alerting is not the job of a virtualization system like Proxmox.

In an ideal world you would not do that no. But most of us Homelabbers have limited resources so you have to be pragmatic. If you have a minimalistic setup (1 host, a couple of guest) using a basic monitoring on/for the host itself could be the solution.

I use Zabbix for that.

Yeah, Zabbix is great. I started with it a while ago but it has quite a steap learning curve (my opinion) and I haven't had the time to set it up properly. Although for a simple basic monitoring it wouldn't be my first choice.

 

[aaron]

For ZFS related email notifications, checkout the section on the ZFS event daemon (ZED) https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_activate_e_mail_notification

 

[Dunuin]

For ZFS related email notifications, checkout the section on the ZFS event daemon (ZED) https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_activate_e_mail_notification

For me personally it was harder to setup zed than setting up zabbix. Sure, setting up zed is super simple. But the hard part was to setup postfix as a relay first so zed is able to actually send emails.