[Request] Disk Temp & Dataset Free Space

[SeiyaGame]

Hi, thanks for your work! It's really insane I was wondering if it was possible to add the temperature of the disks ? I tried on my side and without success, the names of the disks are weird: Also, if possible, I'd like to have the free space of my dataset/pool (To see the evolution of my storage space)

Thanks 👍

[mandrepont]

I had started working on the disk temp side. From what I can tell it is the disk serial number. Here is a mapping I think can be used.

################################################
# Smart log mapping
################################################

- match: "truenas.*.smart_log_smart.disktemp.*.*"
  name: "disk_temp"
  labels:
    job: "truenas"
    instance: "${1}"
    disk_serial: "${2}"

[Supporterino]

Hey there, was looking for disk usage and temp myself. Haven't checked out the smart metrics yet. Need to pull a new set of raw metrics since my test sample was pretty small. Thanks for your findings already. Today was pretty busy for me. But I will get back to working on it tomorrow will take a closer look at your leads and figure something out. Hopefully with good metrics and some graphs on the dashboard.

[Supporterino]

BTW any PRs are totally welcome if you figured it out :)

[mandrepont]

I have been looking through all of the lines I got and don’t see any replacement for dataset usage metrics. That is a disappointment if true. It has been helpful to run the exporter in debug mode for a day and search via Loki and grafana. I’ll keep looking, but I don’t think it’s included in this release.

[Supporterino]

While converting all lines I looked a lot at netdata documentation and netdata is capable of reporting the usage state of zfs pools. I got the input of iXSystem via Reddit that they are looking forward to contributions regarding metrics and exporting them in different formats. So once I am happy with this project and converting the graphite metrics I am going to take a look at the middleware code of truenas since I am also familiar with python.

[mandrepont]

Looking further into the iXSystem side of things. It looks like we can collect disk space from mount path.

I think it would require adding a new plugin to https://github.com/truenas/middleware/blob/stable/cobia/src/middlewared/middlewared/etc_files/netdata/netdata.conf.mako

This one in particular I believe. https://learn.netdata.cloud/docs/data-collection/linux-systems/disk-space

[SeiyaGame]

Looking further into the iXSystem side of things. It looks like we can collect disk space from mount path.

I think it would require adding a new plugin to https://github.com/truenas/middleware/blob/stable/cobia/src/middlewared/middlewared/etc_files/netdata/netdata.conf.mako

This one in particular I believe. https://learn.netdata.cloud/docs/data-collection/linux-systems/disk-space

I've already tried on my own without success The documentation is poor, there is nothing well explained

I modified the /etc/netdata/netdata.conf configuration to activate the plugin:

...
[plugins]
        proc = yes
+       diskspace = yes
        cgroups = no
        tc = no
        idlejitter = no
        perf = no
        apps = no
        nfacct = no
        netdata monitoring = no # We want to disable netdata's agent stats
...

I've tried adding these parameters to the end of the file but it doesn't work, I don't know how to do it ...

[plugin:proc:diskspace:/mnt/*]
#or
[plugin:proc:diskspace:/]
#or
[plugin:proc:diskspace:/mnt]

[Supporterino]

The adjustment of the TrueNAS Netdata and adding new exporters will take place in the new year for me since I am not with my machines and not willing to tinker on them remotely. But thanks for all your input.

The mapping for S.M.A.R.T data and other stuff I find will come soon (tomorrow probably)

[Supporterino]

@mandrepont I found out that I don't even have disk temp in truenas must have been since the upgrade to 23.10. Gonna figure that out first since that is the reason why I missed the metrics. I don't have them :D Gonna keep you guys posted.

[mandrepont]

Weird I was seeing them in my logs and now I am not... On TrueNASCore I got them without even running smart service. Though it seems like it is linked to that service completely.

[Supporterino]

Found an ongoing issue on TrueNAS scale that smart metrics are somewhat broken. Till now there is no fix yet. And since I can't yet find the direct relation to netdata metrics I put the issue on hold. But once I get smart metrics back on my machine I will add them.

[Supporterino]

Temperatures are implemented in https://github.com/Supporterino/truenas-graphite-to-prometheus/releases/tag/v1.1.0 but only with their serial at the moment. Gonna try to map them to their disk names later. Usage is still not exported by TrueNAS

[brantje]

Query to show the disks in a graph: {__name__=~"truenas_truenas_smart_log_smart_disktemp.+"}

Edit: Would it be possible to have de drive serials as label? eg: "truenas_truenas_smart_log_smart_disktemp{sn=xxxxx}

[Supporterino]

You are on an outdated version the metric is names disk_temperature with the serial in the serial label

[Supporterino]

Closing issue now since disk usage isn't reported by truenas anymore