img monitoring proxmox disk usage for home assistant vms

Monitoring Proxmox disk usage for Home Assistant VMs

Troubleshooting Proxmox disk usage for Home Assistant VMs needs a methodical approach. I’ll show what to look for, the exact commands to run, and how to interpret outputs. The usual culprit is the qemu guest agent, but there are a few gotchas with Home Assistant images and LXC containers. Read the commands, run them, and compare expected vs actual.

What you see

Disk usage display in Proxmox UI

You open the Proxmox web UI and the VM shows 0% under disk usage. The Hardware column lists the virtual disk size correctly, for example 32 GiB, but the adjacent usage bar reads 0% or 0.0%. That is the symptom.

Home Assistant VM status

If the VM runs Home Assistant OS (the appliance image), the guest may not provide filesystem stats to the host. If the VM runs Debian/Ubuntu with Home Assistant Supervised or Core in a VM, the guest can run the qemu-guest-agent and report usage. Check the VM type before chasing Proxmox settings.

AdGuard Home LXC server overview

LXC containers normally show disk usage because they share the host filesystem. The UI will report usage derived from the host dataset. If the LXC shows correct disk space while the Home Assistant VM shows 0%, that points at a VM-side reporting problem rather than host storage corruption.

Where it happens

Proxmox host configuration

On the Proxmox host, the VM option “Use QEMU Guest Agent” must be enabled for Proxmox to ask the guest for filesystem info. That option is in the VM Options pane. If it is disabled, the host will not query the guest for disk metrics.

Virtual machine settings

Inside the VM’s Hardware tab the disk is present. Proxmox shows the virtual size (for example 32G). That value is not the used space. The used percentage comes from the guest agent call get-fsinfo. If the guest agent is missing, the UI shows 0%.

Container disk space allocation

LXC containers do not need a qemu agent. Proxmox reads their usage directly from the host dataset. If the LXC (AdGuard Home) reports normal values while the VM reports 0%, treat the VM as the problem.

Find the cause

Guest agent installation

On a standard Linux VM check the qemu-guest-agent service. Run these inside the VM:
bash

Check service status (expected: active/running)

sudo systemctl status qemu-guest-agent

Check process (expected: a running qemu-guest-agent process)

ps aux | grep -i qemu-guest-agent

Expected: service active, process present. Actual: service not found or inactive. If absent, Proxmox cannot get filesystem stats.

On the Proxmox host run:
bash

Ask the guest for filesystem info (replace 100 with your VMID)

qm agent 100 get-fsinfo

Expected: JSON-like output listing mountpoints, total and used bytes. Example:
{“mountpoints”:[{“device”:”/dev/sda1″,”mountpoint”:”/”,”total”:32000000000,”used”:8000000000}]}
Actual: empty response, error, or nothing. If get-fsinfo is empty but the guest agent is running, check permissions and version mismatches.

Compatibility checks

Some appliance images do not include or support the qemu-guest-agent. Home Assistant OS is an appliance image and does not expose a package manager to install services in the usual way. If the VM uses Home Assistant OS, the guest agent is often unavailable and Proxmox will show 0% disk usage for that VM.

Disk usage reporting issues

Proxmox queries guests via the qemu-guest-agent. The host relies on that data for the UI. There are also known cases where get-fsinfo returns values but the Proxmox GUI does not reflect them due to a UI bug or permission problem. See the Proxmox guest agent documentation and forum thread describing the GUI not updating when the agent returns filesystem info [Proxmox qemu-guest-agent wiki] and [Proxmox forum: disk info not displayed].
https://pve.proxmox.com/wiki/Qemu-guest-agent
https://forum.proxmox.com/threads/disk-usage-information-not-displayed-for-specific-vm-in-proxmox-gui.148118/

Fix

Updating Proxmox tools

Make sure Proxmox is up to date and that the node has the correct API permissions. On the host:
bash
apt update && apt full-upgrade

If Proxmox has a known UI bug, an upgrade often fixes it. For monitoring tools that rely on the API, verify any external permission changes after upgrades.

Configuring guest agent

If the VM is a Debian/Ubuntu guest, install and enable the guest agent:
bash

inside the VM

sudo apt update
sudo apt install qemu-guest-agent
sudo systemctl enable –now qemu-guest-agent

Then on the host enable the VM option “Use QEMU Guest Agent” and reboot the VM. Re-run:
bash
qm agent 100 get-fsinfo

Expected: mountpoint list with total and used values.

If the VM runs Home Assistant OS and there is no supported guest agent, accept that Proxmox will not show internal disk percentages. Instead, monitor disk usage from inside Home Assistant. For example, use the Home Assistant supervisor or system information add-on to see disk usage, or enable an exporter that reports metrics to Prometheus.

If get-fsinfo returns values but the UI still shows 0%, try removing and re-adding the node from any external monitoring or management tools and check Proxmox logs for permission errors. Some users fixed the issue by re-adding the node to their monitoring tool or by ensuring the Proxmox API has the node permissions required to read guest agent data.

Verifying storage settings

If the disk is a raw file or qcow2 image, confirm that image size and actual file usage are sensible:
bash

on the host, check image file size

ls -lh /var/lib/vz/images/100/vm-100-disk-1.qcow2
qemu-img info /var/lib/vz/images/100/vm-100-disk-1.qcow2

That shows the virtual size and actual file size. It is not a replacement for in-guest df output, but it helps rule out corrupted images.

Check it’s fixed

Monitoring disk usage

After enabling and starting qemu-guest-agent, run:
bash
qm agent 100 get-fsinfo

Expected: numeric totals and used bytes for each mountpoint. Example snippet:
{“mountpoints”:[{“device”:”/dev/sda1″,”mountpoint”:”/”,”total”:32000000000,”used”:8200000000}]}
If that appears, the Proxmox UI should update within a minute.

Running diagnostic commands

Inside the VM confirm filesystem stats:
bash
df -h /

Expected: shows used and available space that matches the guest agent numbers (used bytes / total). If df shows 8.2G used and get-fsinfo shows roughly 8.2G in bytes, the connection is correct.

On the host, if the command returns an error like “guest agent did not respond”, check VM Options, enable the guest agent, and reboot the guest.

Confirming resolution with Proxmox UI

Refresh the Proxmox web UI. The disk usage column should now show a percentage and the usage bar should correspond to the guest numbers. If it still reads 0%, check the host’s syslog and Proxmox logs for permission or API errors and confirm the node is not running an older UI with a known bug.

Final takeaways: if the VM runs a normal Linux guest, install qemu-guest-agent, enable it in VM options, and verify with qm agent get-fsinfo and df -h. If the VM runs Home Assistant OS, the gues agent may not be available, so monitor disk space from inside Home Assistant or use a metrics exporter.

Leave a Reply

Your email address will not be published. Required fields are marked *

Prev
Weekly Tech Digest | 04 Jan 2026
weekly tech digest

Weekly Tech Digest | 04 Jan 2026

Stay updated with the latest in tech!

Next
Best practices for Proxmox Backup Server snapshot management
img best practices for proxmox backup server snapshot management

Best practices for Proxmox Backup Server snapshot management

Proxmox Backup Server snapshots can yield corrupt metadata if taken while PBS is

You May Also Like