Support for additional disks / partitions

[henrygd]

Update: Initial support has been added in 0.3.0. See the readme section for instructions.

This is a much requested feature (#2, #3, #23) so I want to direct discussion here.

Any opinions on the best way to implement and display would be appreciated. I'd like it to be as straightforward as possible.

Should we just have an environment variable EXTRA_FILESYSTEMS where you would supply a comma separated list?

Then they would be laid out separately like this mock up?

[bozzio83]

That sounds great to me!

[chachoCL]

🙏

[crypt0rr]

Yes, please! 🚀

[JamsRepos]

Yeah, good implementation. I saw it as working as all in one array but this makes a little more sense as allows you to monitor the host and arrays seperately.

Although, perhaps allowing multi file systems for both? Then if people have a pool which is their primary filesystem they can pool those together than the extra for any other system devices?

[dnburgess]

This looks amazing!!

[henrygd]

Thanks David, loved the video!

[dnburgess]

So glad you liked it!!

[SoNotMC]

Yes I would love this, means I can monitor all my different file systems in one place!

Also could there be a posbility of adding GPU support? Since I host a jellyfin container and being able to see GPU utilisation would be great.

[henrygd]

GPU stats are not supported by psutil or gopsutil, and apparently it's not an easy thing to add. So not likely.

GPU temp is a possibility though.

[SoNotMC]

Sounds good. nvidia-smi is a thing but im unsure on how you would get information from it due to its formatting.

[palijn]

The proposed solution looks great to me.
IF it makes sense to you (like, if the code is easier to reuse), displaying a single graph with aggregated data like the Docker usage would work pretty well for me too.

Any guess as to when you would be able to deliver this feature ? :)

[henrygd]

Hopefully by the end of the month.

[pabloalcantara]

Great. This way it´ll be best

[ctrlaltdelete007]

looks great!!!

[henrygd]

Working on implementing this now.

The EXTRA_FILESYSTEMS environment variable works for the binary version of the agent.

However, I can't find a way to get disk size / usage in Docker unless a folder from the device is mounted in the container.

So for the Docker agent, you would NOT use EXTRA_FILESYSTEMS, but rather do something like below. The agent would then pick up any device that has a mountpoint in /extra-filesystems.

volumes:
  - /mnt/disk1/.beszel:/extra-filesystems/disk1:ro
  - /mnt/disk2/.beszel:/extra-filesystems/disk2:ro
  - /dev/mmcblk0/.beszel:/extra-filesystems/sd-card:ro

Please let me know if this method would present issues for anyone. Or if you have an idea for a better workaround.

[henrygd]

I haven't tested, but you should be able to combine FILESYSTEM and EXTRA_FILESYSTEMS now to override which partition is picked up as root. However, you would get duplicate charts because it would still be picked up as an extra fs.

Maybe it would make sense to add a rule where you can override root by naming the mounted folder rootfs?

volumes:
  - /mnt/disk1/.beszel:/extra-filesystems/rootfs:ro
  # or maybe outside of /extra-filesystems is better?
  - /mnt/disk1/.beszel:/rootfs:ro

Edit: Actually you may be able to do /mnt/disk1/.beszel:/rootfs:ro currently and set FILESYSTEM to use that without needing to mount in /extra-filesystems.

[JamsRepos]

Yeah, that would make a lot more sense, as i'm sure a lot of Unraid users would want their /mnt/cache to be their /rootfs

[henrygd]

@JamsRepos I have trouble understanding the Unraid setup since I have no experience with it.

Referencing your df output in the Unraid discussion, it looks like your system is using /var/lib/docker for calculating disk usage, since that's the actual storage location for the container, but you'd prefer to use /mnt/cache, i.e., partition nvme0n1p1. So ideally it would display 30% usage instead of 19%. Do I have that right?

[JamsRepos]

@henrygd That would be correct yeah. Obviously my Unraid setup may differ from others but the ability to set this to the path of your choice would be great, although in most I would assume others would also use their cache.

But yeah, it appears to be using /var/lib/docker and I'd like to change to the nvme0n1p1 as per my command output.

[henrygd]

@JamsRepos I think this is possible now in 0.4.0 by doing this:

environment:
  FILESYSTEM: nvme0n1p1
volumes:
  - /mnt/cache/.beszel:/extra-filesystems/cache:ro

FILESYSTEM specifies which partition should be considered root, and mounting a folder gives the agent access to check the usage.

You can also try using "/extra-filesystems/cache" as the FILESYSTEM value.

If that doesn't work, let me know what the agent logs.

[death4444]

Technically unaware, demand +1，Looking forward to it.

[DominikBritz]

I use the beszel binary and have it running as systemd service. How do I add the environment variable EXTRA_FILESYSTEMS ?

[scriptzteam]

Something like this one?

beszel/supplemental/guides/systemd.md

Line 113 in d2284c3

Environment="PORT={PASTE_YOUR_PORT_HERE}"

Environment="EXTRA_FILESYSTEMS=sdb"

:)

[DominikBritz]

Perfect, thank you

[Killerbeast01]

Any posibility to change the displayed name for the disk? like i have a lxc container with three disks, and all just get printed like "DataGrave-vm--100--disk--0", "DataGrave-vm--100--disk--1", etc...
Bind via volume mount for the docker agent:

volumes:
      - /var/run/docker.sock:/var/run/docker.sock:ro
      - /mnt/dataGrave/.beszel:/extra-filesystems/dataGrave:ro
      - /mnt/dataGrave/Media/.beszel:/extra-filesystems/Media:ro
      - /mnt/dataGrave/Backups/.beszel:/extra-filesystems/Backups:ro

[henrygd]

See #379. We'll add a way to do this eventually either through a config file or maybe delimiter in the folder name.