Varqeni Help

Resources

Updated 11 November 2026

Resources is the inventory view: the Proxmox host and its virtual machines, containers, templates and storage, the Kubernetes nodes, object storage and the database, read from what Varqeni already watches. It is laid out the way Proxmox and Grafana lay out an estate, so it reads at a glance.

AvailabilityAvailable to everyone with early access to Varqeni.

Steps

  • The tree: on the left, every item under its host, the way Proxmox lists them: each guest by its number and name with dots in its tag colours, then templates and storage. Switch it to group by type or by tag. On a phone it becomes a list to pick from.
  • Resource allocation: one row per item with its vCPUs, CPU used, memory given and used, disk, uptime, tags and Varqeni's own health light with the count of open findings. Sort by any column, filter by name, number, type, tag or state, and choose the columns. Green, amber and red follow Varqeni's thresholds.
  • Two memory figures: Memory used is what Proxmox sees, and it counts a guest's page cache as used, so a busy guest often looks nearly full. In guest is the guest's own figure from its node exporter, without cache: the real pressure. Varqeni colours a guest by the second where it has it.
  • The host and its guests: host CPU and memory over the time bar's range, with the amber and red bands behind them and gauges for now; every guest's CPU and memory as one coloured line each, with a table of its highest, average and latest value; storage gauges and space allocation; container disk; network received above the line and sent below; disk read and written.
  • One item: choose a row or a tree item to open its page. Summary shows its status, CPU, both memory figures, swap, disk, boot disk, network, HA, start at boot and backup coverage, with charts you can switch between average and maximum. Findings lists what Varqeni sees now and the incidents that named it. Runs here lists the pipelines and services on it. Actions shows the fixes an open incident recommends for it, with the exact command, and its task history. Monitoring says what watches it and what is missing. Settings shows its thresholds and what Varqeni may do to it.
  • Fixes: a fix appears only while an open incident recommends it for that item, and approving it goes through the same checks and named approval as anywhere else in Varqeni. Resources never changes anything by itself.
  • The real tools: Open in Proxmox and Console open the guest in Proxmox itself; Grafana opens the estate's dashboard where it is connected. Explain asks PelQi about the item with its live numbers.
  • Keys: / filters the table, the arrow keys move between rows and Enter opens one. Each page has its own address, so a link opens on the same item and section.
  • Refresh and the past: the page refreshes every 30 seconds unless you change it. Charts follow the time bar; the table always shows now, because Varqeni's archive keeps each system's health, not every guest's usage.

What you should see: Numbers come from the Proxmox exporter, the node exporter in each guest, kube-state-metrics and MinIO's own metrics, through Varqeni's read-only Prometheus login. A guest without a node exporter shows only what Proxmox sees, and Monitoring shows how to add one.

What it can and cannot do

Can

  • List every machine with its allocation and use, show Proxmox's memory beside the guest's own, chart the host and every guest over the time bar, and open each item's own page with its findings, workloads, fixes and coverage.

Cannot

  • Change a machine from this view, or show past usage for every guest: the archive keeps each system's health, not per-guest numbers.

If it does not work

  • Not watched from inside: the guests named have no node exporter, so their memory and disk come from Proxmox only, and Proxmox counts page cache as used. Add a node exporter to the guest (its Monitoring tab says how) to see its own figures.
  • A link opens the list, not the item: the item may have been removed or renamed in Proxmox. Pick it again from the tree.

Still stuck? Contact support or email info@fortisdatagroup.com.