Disk Health in Metric Server stream

[proxdox]

Hi,
First, I am very thankful for this wonderful software and community.

I have the PVE metrics stream configured to an InfluxDB. All is well and I love the detailed info. One detail that would be very useful is Disk Health.
For example, Disks' S.M.A.R.T and Wearout info. When I search there are some scripts and call for SmartCTL, Telegraf plugin to be installed and so on.
I'd prefer to not install anything other than what PVE does by itself and tried to pull metrics through a container instead but you already know that provides only limited info and not the SMART and wearout (unless I don't how to).

Are there any plans for this info to be included in the Metrics Server stream? It would just simplify so much!

Thanks!

 

[dcsapak]

Are there any plans for this info to be included in the Metrics Server stream? It would just simplify so much!

not that i'm aware, but you can always open a feature request on https://bugzilla.proxmox.com (no promises though)

in general, there speaks nothing against installing things like telegraf on the proxmox host directly, as long as you know what the packages do and can handle any problems or issues with it yourself

you could also pull the data via the pve api inside a container that sends it to your monitoring

 

[proxdox]

in general, there speaks nothing against installing things like telegraf on the proxmox host directly, as long as you know what the packages do and can handle any problems or issues with it yourself

True, but call it experience or paranoia,I generally try to avoid this due to instabilities and issues that could get introduced due to those said packages.

you could also pull the data via the pve api inside a container that sends it to your monitoring

To test, Installed lm-sensors, telegraf in a container and was able to pull some data but limited to virtual disk assigned.
Perhaps full access requires a privileged container, something I am trying to avoid, hence the request for this info to be included in Metric Server reporting.


Currently, how is the disk health monitored in a data center environment with many PVE nodes and disks?

Thanks!

 

[dcsapak]

Currently, how is the disk health monitored in a data center environment with many PVE nodes and disks?

Usually there are other monitoring solutions in place, either installed on the host directly or some don't need anything special installed, e.g. checkmk (these can often connect via ssh and check the system with standard tools or via the api)

 

[proxdox]

Usually there are other monitoring solutions in place, either installed on the host directly or some don't need anything special installed, e.g. checkmk (these can often connect via ssh and check the system with standard tools or via the api)

Thanks!