Browse documentation
Reading a hypervisor
The hosts list
The left rail filters by view and by platform:
| View | Shows |
|---|---|
| All hosts | Everything. |
| Needs attention | Hosts with something wrong. |
| High CPU (>70%) | Hosts under load. |
| Low storage | Hosts running out of room. |
| Maintenance mode | Hosts deliberately taken out of service. |
| Standalone hosts | Hosts not in a cluster. Remember every Proxmox node reads as standalone — see Supported platforms. |
Below the filters, a capacity summary shows how much of the estate's memory and storage is in use — the figure to watch before promising anyone a new VM.
A single host
The host page opens with four KPIs — CPU, Memory, and capacity figures — each noting when it was last polled, and coloured as they approach their limits.
Below that:
- Host resource usage — CPU and memory over time.
- Storage — the host's volumes. A host that hasn't reported disk usage says so rather than showing zero.
- Virtual machines — the guests running on this host.
Notes & tips
- Memory is usually the binding constraint on a virtualisation host, not CPU. The memory KPI and the estate capacity bar are the two numbers worth watching.
- A host in maintenance mode is filtered out of the normal views deliberately — check that list before wondering where a host went.