Disk space/stats unavailable randomly

[agit8or1]

All vms have qemu guest agent install and working, yet some vms get this error:

Guest details unavailable
Disk stats unavailable. Guest agent may not be installed.

Install and run the qemu-guest-agent inside this VM so Pulse can surface OS, network, and filesystem details.

It's intermittant also. Sometime they work, sometimes they dont

[rcourtman]

Hi @agit8or1.

Pulse raises the Guest details unavailable and Disk stats unavailable messages when Proxmox returns an error to the guest agent calls (guest info and guest get fsinfo). That can happen when the agent service or the virtio serial channel stalls for a moment, and the warning clears only after the next successful poll.

Please gather the following:

1. Inside an affected VM while the warning is present run systemctl status qemu-guest-agent and journalctl -u qemu-guest-agent and note any restarts or transport errors around the same time.
2. On the Proxmox node run qm agent <vmid> ping and qm agent <vmid> get-fsinfo a few times. Note whether any call fails during the same window.
3. Confirm the VM configuration still has the Qemu Agent option enabled. Note whether backups or snapshots start just before the warning appears.

Export diagnostics from the application by opening Settings → Diagnostics, running the diagnostics, then choosing Export for GitHub. Review the JSON to confirm the redaction is acceptable, then attach it here along with the items above. That export shows which VMs Pulse recorded as failing to return filesystem data, which will help confirm whether the host call is timing out.

Also share the relevant lines from /var/log/pulse/pulse.log that mention the guest agent around the warning time and list the VM IDs that show the problem most often.

I will compare those details once they arrive.

[agit8or1]

So it's completely random and happens to ALL vms (55 Total). It comes and goes. I have been unable to pinpoint a cause or time. All vms have qemu guest agent and they all work. I have double checked them via Proxmox dashboard. I'll work on tracking more info down. Maybe a polling timeout which also could be related to all of the vms intermittantly not being displayed (This happens less frequently however). Checked all the obvious, network issues, etc. Nothing found

pulse-diagnostics-sanitized-2025-10-23.json

[agit8or1]

taken a few minutes apart