Proxmox Node Went Grey but the VMs Kept Running: What Actually Breaks?

A grey question-mark node does not always mean the host is down. In one real case, restarting pvestatd brought the node back while the deeper cause remained harder to pin down.

A Proxmox node suddenly turns grey in the web interface. Question marks appear beside it. The VMs and containers are still running normally. A reboot fixes everything, but only temporarily.

The node may be alive while the status layer is dead

The original setup had been showing this behavior for a long time, across much of the Proxmox 8.x line and again on 9.2.4. Sometimes the problem appeared after a day. Sometimes the cluster ran for a month before one node went grey.

The important clue was that the workloads kept running.

That immediately suggests a distinction between the hypervisor itself and the services Proxmox uses to collect and display status information.

The highest-rated response asked one simple question: is pvestatd still running?

That turned out to be the quickest win.