Temperature

[Johannes S]

And there is already code in that direction, submitted for inclusion: https://bugzilla.proxmox.com/show_bug.cgi?id=208#c23
It's a halfway, but it would be already a good step. But it doesn't get enough attention.

Which code are you meaning? https://github.com/Meliox/PVE-mods doesn't address Andreas Steinels (aka @LnxBil ) argument that typical servers have a lot of temperature sensors which would clutter the UI and for monitoring it's best to setup a dedicated monitoring software anyhow. For inclusion the author of the mod would have to sign a contributor agreement. In corportate environments Prometheus, Zabbix or a nagios variant (Icinga2, checkmk etc ) are often used for monitoring, for the typical homelab something like Pulse (https://github.com/rcourtman/Pulse ) should fit your needs. Pulse has temperature monitoring btw: https://github.com/rcourtman/Pulse/blob/main/docs/TEMPERATURE_MONITORING.md

 

[templar]

Which code are you meaning?

Sorry for asking, but have you followed the link I gave? It leads direct to the code's author's comment with his explanation and links to the submitted patches. If he did sign the agreement or not - I don't know, you can ask him yourself.

 

[Johannes S]

Sorry for asking, but have you followed the link I gave? It leads direct to the code's author's comment with his explanation and links to the submitted patches. If he did sign the agreement or not - I don't know, you can ask him yourself.

Sorry, missed that comment, thanks for the hint. . i think adding this information to the API is a good thing since it makes it easier to implement this functionality in third-party-monitoring (e.G. pulse, ProxmoxVE-specifig plugins for Icinga2/Zabbix/Prometheus exporters), so in that regard these patches are a good enhancement and will hopefully included. I still think that it's these information doesn't belong in the main dashboard because this would clutter the UI for the reasons explained by Andreas Steinel and others.

 

[templar]

adding this information to the API is a good thing since it makes it easier to implement this functionality in third-party-monitoring

Exact my thoughts too. I personally use Zabbix and just wanted to omit installing of its agent. Zabbix already has template for reading from PVE API, and PVE API already has S.M.A.R.T. data. And if it would return also sensors data I would be totally happy.

I still think that it's these information doesn't belong in the main dashboard because this would clutter the UI

And I can agree with you here too, I can't deny it

 

[DerekG]

Exact my thoughts too. I personally use Zabbix and just wanted to omit installing of its agent. Zabbix already has template for reading from PVE API, and PVE API already has S.M.A.R.T. data. And if it would return also sensors data I would be totally happy.

As I understand it the sensors data is already available via Proxmox api, Pulse Monitor manages to pull the info for the CPU and NVMe at least.

 

[Johannes S]

As I understand it the sensors data is already available via Proxmox api, Pulse Monitor manages to pull the info for the CPU and NVMe at least.

Yes but not for temperature, the pulse featore for monitoring temperatures it uses a specific agent together with lm-sensors see: https://github.com/rcourtman/Pulse/blob/main/internal/sensors/collector.go , https://github.com/rcourtman/Pulse/blob/main/docs/TEMPERATURE_MONITORING.md and https://github.com/rcourtman/Pulse/blob/main/ARCHITECTURE.md

 

[patch]

I still think that it's these information doesn't belong in the main dashboard because this would clutter the UI

Agree it belongs in a "Hardware" tab mirroring what is already there for VM's. On the dashboard just showing the highest current or range of all sensors would be optimal.

 

[Johannes S]

Agree it belongs in a "Hardware" tab mirroring what is already there for VM's

The hardware tabwould also be cluttered as explained by Andreas Steinel in the bug ticket:

Normally, servers have more than 20 temperature sensors (including the disks), so ONE display can never be sufficient. You should just look into external monitoring which is better suited to do what you want ... MONITORING, not just a display
https://bugzilla.proxmox.com/show_bug.cgi?id=208#c16

And since temperature monitoring is easily available by using a real monitoring software it's not even needed.