Proxmox VE Custom

Enter the address

• Host
• Password or token value
• Token name
• Port
• Realm
• Username
• Verify SSL certificate

Also asked when changing the connection.

Expose

• Nodes
• Virtual Machines (QEMU)
• Linux Containers (LXC)
• Storages
• Monitor physical disksRead the model, size, temperature, wearout and SMART health of each node's physical disks as a device of its own. If you spin disks down, switch this off - the SMART read wakes them on every poll otherwise.
• Track everything automaticallyTrack every node, virtual machine, container and storage the credentials can see, and follow the cluster from then on: whatever is added is picked up, whatever is removed is dropped. The selection above is ignored while this is on. Templates are never tracked.
• Entity id schemeStandard: Home Assistant's device then name, qemu_win11_108_cpu_used. Extended: prefix, kind, id, name, item - pve_qemu_108_win11_cpu_used - so a search or a recorder filter catches everything by the prefix and a list sorts by id. Applies to entities created from now on; ids that exist keep them.
• Prefix for the extended schemeThe first part of every entity id under the extended scheme; pve unless you choose otherwise. Not used by the standard scheme.

Reauthenticate Integration

• Password or token value
• Username
• Token name
• Realm

After setup

• Password or token value
• Token name
• Realm
• Username
• Verify SSL certificate
• UsernameLeave empty (along with the fields below) to disable the Arm/Disarm HA buttons and the HA managed sensor.
• Token name
• Password or token valueNeeds Sys.Console (arm/disarm) and Sys.Audit (HA resource list) on '/' — not the scoped permissions recommended for the main integration user.
• Realm
• Linux Containers (LXC)
• Nodes
• Virtual Machines (QEMU)
• Storages
• Track everything automaticallyTrack every node, virtual machine, container and storage the credentials can see, and follow the cluster from then on: whatever is added is picked up, whatever is removed is dropped. The selection above is ignored while this is on. Templates are never tracked.
• Monitor physical disksRead the model, size, temperature, wearout and SMART health of each node's physical disks as a device of its own. If you spin disks down, switch this off - the SMART read wakes them on every poll otherwise.
• Monitor failed tasksTrack and report failed tasks from the last 24 hours on the selected nodes, as a sensor. Read every five minutes.
• Monitor package updatesRead the pending package upgrades of each node into an update entity and a count sensor. Needs Sys.Modify on the node; switch this off to run without that privilege - or without the entities.
• Polling intervalHow often nodes, guests, storage and backups are read. 60 seconds unless changed; Proxmox itself refreshes guest figures about every 10 seconds. Certificates, subscriptions and Ceph are read hourly, failed tasks every five minutes, whatever is set here.
• Guest file path to monitor (QEMU, requires guest agent)Absolute path of a file inside each tracked QEMU guest to read via the guest agent and expose as a sensor. Leave empty to disable. Requires the QEMU guest agent to be running in the guest.
• Backup storage for the backup buttonsWhere the "Back up now" and "Back up all" buttons write their backups (snapshot mode). The buttons exist only while a storage is picked here and the credentials hold VM.Backup. Storages that accept backups are listed; any other name can be typed in.
• Entity id schemeStandard: Home Assistant's device then name, qemu_win11_108_cpu_used. Extended: prefix, kind, id, name, item - pve_qemu_108_win11_cpu_used - so a search or a recorder filter catches everything by the prefix and a list sorts by id. Applies to entities created from now on; ids that exist keep them.
• Prefix for the extended schemeThe first part of every entity id under the extended scheme; pve unless you choose otherwise. Not used by the standard scheme.