Proxmox Extended Sensors for PVE and PBS inside Home Assistant

:puzzle_piece: New Integration: Proxmox Extended Sensors

:link: GitHub Repository

Advanced monitoring and maintenance actions for PVE and PBS inside Home Assistant

Hello everyone,

I’d like to introduce Proxmox Extended Sensors, an integration designed to provide a more complete, structured, and powerful experience when working with Proxmox VE and Proxmox Backup Server from within Home Assistant.

This project was created to complement and extend the existing capabilities, adding advanced metrics, detailed sensors, and new features aimed at improving daily administration of Proxmox environments.


:magnifying_glass_tilted_left: What does this integration offer?

:check_mark: Advanced monitoring for PVE

  • Sensors for CPU, RAM, storage, node status, and VM/CT status
  • Network RX/TX sensors for virtual machines and containers
  • Clean device grouping for a clear and organized UI

:check_mark: Direct actions on VMs and CTs

Buttons for:

  • Start
  • Shutdown
  • Stop
  • Reboot
  • Reset
  • Suspend / Resume
  • Hibernate (when supported)

All accessible directly from Home Assistant without opening the Proxmox interface.


:shield: PBS integration and maintenance tools

:check_mark: New maintenance buttons

Trigger PBS maintenance tasks directly from Home Assistant:

  • Garbage Collect
  • Prune
  • Verify
  • Sync

:check_mark: New “Last Action” sensor

A local sensor that displays the last maintenance action performed.

  • Initial state: “Waiting”
  • Instantly updated when any maintenance button is pressed
  • Fully local and independent from the coordinator
  • Stable and persistent across restarts

:check_mark: Organized by datastore

Each datastore has its own device grouping, keeping everything clean and easy to navigate.


:broom: Internal improvements and cleanup

  • Full refactor of the PBS module
  • Removal of obsolete entities
  • Cleaner, more maintainable codebase
  • Improved compatibility with Home Assistant

:motorway: Roadmap

  • Custom Lovelace card templates
  • Submission to HACS as an official repository
  • Advanced PBS maintenance sensors (progress, duration, logs)
  • Optional advanced PBS dashboard
  • Create VM/CT backups directly from PVE via Home Assistant

If you’d like to try it, share feedback, or suggest improvements, you’re more than welcome.
Every idea helps refine the integration and push it even further.

Thanks, and I hope you find it useful.

4 Likes

Release v1.3.0 — Proxmox Enhanced Backup & Hardware Compatibility

This update introduces the long-awaited manual backup service directly from Home Assistant, alongside critical optimizations for hardware monitoring and PBS management.

:rocket: New Feature: Manual Backups with HA Identification

You can now trigger backups for your VMs and Containers directly from Home Assistant.

  • Custom Action: Added the proxmox_sensors.create_vzdump_backup service.
  • Automatic Tagging: Backups triggered from HA are automatically labeled as ID | HA Backup to easily distinguish them from Proxmox scheduled tasks.
  • Full Deduplication: Fully compatible with PBS deduplication, maintaining storage efficiency.

:globe_with_meridians: Remote Backup Support (Tuxis, Offsite PBS, etc.)

This version has been successfully tested with remote storages. You can now trigger backups to offsite locations (like Tuxis or any remote Proxmox Backup Server) directly from Home Assistant.

  • How it works: As long as the remote storage is already configured in your Proxmox Node, simply provide its name in the storage field of the action.
  • Smart Performance: Proxmox handles the connection and deduplication. Even for remote backups, only the modified data chunks are transferred, ensuring fast and efficient offsite copies.
  • Verified: Confirmed working with remote PBS providers, including automatic tagging and Gotify notifications if configured in Proxmox.

Example YAML:

action: proxmox_sensors.create_vzdump_backup
data:
  node: your_node_name
  vmid: 100
  storage: YOUR_REMOTE_STORAGE_NAME


New service interface in Home Assistant.


Identified backups as seen in Proxmox Backup Server.


:hammer_and_wrench: Hardware & Compatibility Improvements

  • LM-Sensors Optimization: Improved handling of CPU temperature and fan speed readings across various Intel-based architectures and chipsets.
  • Client Refactoring: Deep cleanup of the internal API (client.py) to improve response times and request stability.
  • Log Cleanup: Removed redundant debug messages in button.py to keep the Home Assistant System Log clean.

:bar_chart: Proxmox Backup Server (PBS) Enhancements

  • Human-Readable Dates: Task and backup timestamps now use a standard format (DD/MM/YYYY HH:MM) instead of Unix timestamps.
  • Datastore Management: Improved handling of multiple datastores and maintenance states (GC, Prune, Verify).

:memo: Update Instructions

  1. Wait for HACS to notify you about the new version v1.3.0.
  2. Click on Update in the HACS dashboard.
  3. Restart Home Assistant to apply the changes.
  4. Verify that the new action appears in Developer Tools > Actions (formerly Services).
1 Like

Looks promising! :slight_smile:

I’ll give it a try.

Have you used GitHub - dougiteixeira/proxmoxve: Proxmox VE Custom Integration Home Assistant before? Would you say, your integration includes all of its features?
I’m currently using it. Unfortunately, it seems there’s no further development going on.

I got it running.
As expected looks very good :slight_smile:

Any idea why the integrations icon is not loading in the UI?

You are querying PBS via API for task information and more.
I have realized that by using webhooks.
I am not a developer, so I cannot say which approach is “better” (more robust, consumes less resources, is easier to set up, …).
Just wanted to give a heads up, that both exists :slight_smile:

"Hi @chris400, thank you so much for your feedback! I’m glad to hear the integration is working well for you and that you like the UI.

Regarding your comments:

API vs Webhooks: You are right, I’m using the Proxmox/PBS API directly. I personally find this approach much more robust and reliable. While webhooks depend on the server “pushing” the data (and can sometimes fail or require extra configuration), querying the API ensures that Home Assistant always has the real-time status exactly when it needs it, without losing sync.

  • Backup Management: One of the main goals of this integration was to go beyond simple monitoring. Currently, you can already trigger individual backups for your CTs and VMs directly from Home Assistant.

  • What’s coming next: I’m finishing a major update. Soon, you will be able to trigger bulk backups (all VMs/CTs at once) with full control over the process: choosing the backup type (Snapshot, Suspend, or Stop) and compression settings** (ZSTD, LZO, GZIP), all from your HA dashboard.

The icon issue is simply because I’m still in the process of submitting the integration to the official Home Assistant “Brands” repository and preparing it for HACS. Stay tuned!

Thanks again for the support!"

1 Like

:rocket: Proxmox Extended Sensors v2.0.0 — Major Release

:united_kingdom: English

This version represents a complete leap forward in functionality, stability, and integration depth.
It includes a full redesign of the sensor system, new control capabilities, full PBS compatibility, and a fully optimized structure for Home Assistant.

:fire: Main Highlights

  • Full compatibility with Proxmox VE 7/8/9 and PBS 3/4
  • New advanced sensor system for hardware, VMs, CTs, disks, and PBS
  • Fully integrated individual and mass backup services
  • Full PBS deduplication compatibility
  • Complete control of nodes, VMs, CTs, and PBS
  • Clean and organized entity/device structure
  • Secure token‑based authentication
  • Translations available in 8 languages
  • Optimized polling and reduced resource usage
  • Fully renewed documentation

:puzzle_piece: Internal Changes

  • Rewritten configuration flow (config_flow)
  • Complete code reorganization for improved stability
  • New endpoints and improved error handling
  • Better compatibility with custom dashboards (Card‑Mod, Mushroom, etc.)

:blue_book: Additional Notes

This version represents a new generation of the integration, built with a professional, scalable approach and prepared for future expansions.


:spain: Español

Esta versión marca un salto completo en funcionalidad, estabilidad y alcance de la integración.
Incluye un rediseño profundo del sistema de sensores, nuevas capacidades de control, compatibilidad total con PBS y una estructura completamente optimizada para Home Assistant.

:fire: Novedades principales

  • Compatibilidad completa con Proxmox VE 7/8/9 y PBS 3/4
  • Nuevo sistema de sensores avanzado para hardware, VMs, CTs, discos y PBS
  • Servicios de backup individuales y masivos totalmente integrados
  • Compatibilidad total con deduplicación PBS
  • Control completo de nodos, VMs, CTs y PBS
  • Estructura limpia de entidades y dispositivos
  • Autenticación segura basada en API Tokens
  • Traducciones completas en 8 idiomas
  • Optimización del polling y reducción del uso de recursos
  • Documentación completamente renovada

:puzzle_piece: Cambios internos

  • Reescritura del flujo de configuración (config_flow)
  • Reorganización completa del código para mayor estabilidad
  • Nuevos endpoints y manejo mejorado de errores
  • Mejor compatibilidad con dashboards personalizados (Card‑Mod, Mushroom, etc.)

:blue_book: Notas adicionales

Esta versión representa una nueva generación de la integración, con un enfoque profesional, escalable y preparado para futuras ampliaciones.

1 Like

:rocket: v2.1.0 – Scalable Architecture & Advanced Backup Control

This version marks a milestone in the project’s maturity.

It’s not just about bug fixes, but a complete consolidation of the architecture and backup system.

The integration is now ready for multiple Proxmox VE and Proxmox Backup Servers, with stable identifiers, automatic migrations, and a much more flexible and controlled backup engine.


:hammer_and_wrench: System Fixes and Improvements

:white_check_mark: Dynamic Entity Naming (Multi-Server)

A unique identifier system has been implemented, allowing for multiple PVE and PBS instances without conflicts.

Before:

sensor.task_task_last_task

Now:

sensor.task_pbs_1_last_task

(The index is automatically assigned per server)

Includes seamless automatic migration:

Your dashboards and automations WILL NOT be broken.

No need to delete entities or reconfigure Home Assistant.


:white_check_mark: Silent Automatic Migration

Existing installations are updated internally without user intervention.

The integration adjusts internal identifiers while keeping the same visible entity_ids, ensuring full compatibility.


:white_check_mark: Power Control Fixed

Fixed the issue that prevented the following actions from working correctly:

  • Start
  • Stop
  • Restart

on certain VM and CT configurations.


:sparkles: New Features

:brain: Physical Memory (DIMM) Monitoring

Sensors have been added to detect the actual RAM modules in the node, including:

  • Speed ​​(MT/s) and configured speed
  • Memory manufacturer and type
  • Physical slot on the motherboard

Ideal for advanced diagnostics and hardware maintenance.


:floppy_disk: Advanced Backup Management

The backup system has been redesigned:

:small_blue_diamond: Selective Backup

You can now specify multiple VMs and CTs directly:

100, 103, 124

or as a YAML list:

guests:
  - 100
  - 103
  - 124

:small_blue_diamond: Sequential Execution

Simple backups are now executed strictly one by one, respecting configurable delays between each operation, preventing overloading the node’s I/O subsystem.

This allows for controlled backups even on limited hardware.


:small_blue_diamond: Bulk Backup Mode

Parallel backup mode with concurrency control is maintained for scenarios where speed is a priority.

Each user can choose between:

  • Sequential (controlled) backups
  • Bulk (parallel) backups

depending on their needs.


:white_check_mark: Compatibility

This version is fully compatible with existing configurations.

No manual action is required after updating.

Existing automations, dashboards, and entities continue to function without changes.


:man_in_lotus_position: Why this version matters

This release does not add any uncontrolled sensors.

It consolidates the project’s foundation:

  • Secure migrations
  • Stable identifiers
  • Controlled backups
  • Real support for multiple servers
  • Ready for future evolution

This version establishes a solid foundation for project growth.


:inbox_tray: PVE Extractor Script Update

If you are using advanced node monitoring (lm-sensors, physical disks, DIMMs, etc.), you need to update the system script on the Proxmox host.

Run the following command on the PVE node:

# Download or update the extractor script
wget https://raw.githubusercontent.com/Javisen/proxmox_sensors/main/scripts/pve-sensors-api.py -O /usr/local/bin/pve-sensors-api.py && chmod +x /usr/local/bin/pve-sensors-api.py
# Reload systemd and enable the service
systemctl daemon-reload
systemctl enable --now pve-sensors

This ensures compatibility with the new sensors introduced in this version.

:warning: Recommended to run after updating the integration.


:heart: Thank you

Thank you to all the users who test the integration, provide feedback, and help improve the project.

1 Like

:rocket: Proxmox Sensors V3 – Coming Soon

After a major rework, V3 is almost ready. This release focuses on cleaner architecture, smarter sensors, and real system insight instead of just raw metrics.


:new_button: New Features

  • IO Wait sensor (node)
  • Load Average sensor (1m + 5m/15m attributes)
  • Node Score sensor (overall node performance)
  • Node Stressed binary sensor (CPU + IO + Load)
  • Node Disk Overloaded binary sensor
  • CPU attributes with cores and per-core usage (CT/VM/Node)
  • Automatic node discovery
  • Manual node selection during setup
  • Wake-on-LAN (WOL) service for PVE and PBS (via MAC)

:wrench: Improvements

  • Clean CT sensor naming (id - name format)
  • Removal of duplicated entity names
  • Simplified and consistent unique_id
  • Improved extra_state_attributes across sensors
  • Normalized units (%, GB, etc.)
  • Accurate CPU usage calculations everywhere
  • Safer data handling and fallbacks
  • Improved PBS task parsing and filtering
  • Better datastore extraction from tasks (worker_id)
  • Verify/Prune task handling improvements
  • Removal of debug logs and cleanup for production
  • Unified temperature sensors (CPU + NVMe)
  • Storage filtering per node

:brick: Architecture

  • Clear separation between sensor and binary_sensor
  • Proper use of device_class
  • Consistent state_class usage
  • Standardized attributes across all entities
  • Ready for multi-node environments

:bullseye: Result

  • Intelligent sensors (not just raw data)
  • Better visibility of real system state
  • Strong foundation for automations
  • Clean, scalable, production-ready integration

:warning: Important:
V3 introduces breaking changes and will require reconfiguring dashboards and entities.


Stay tuned — release coming soon.

:rocket: Proxmox Extended Sensors V3 – Full PVE & PBS Monitoring for Home Assistant

Hi everyone!

I’m happy to share V3 of Proxmox Extended Sensors, a custom integration for Home Assistant that provides deep visibility and control over your Proxmox infrastructure.

This release is a major step forward, especially with full Proxmox Backup Server (PBS) support and improved system insight.


:brain: What makes it different?

Instead of exposing only raw metrics, this integration focuses on actionable information.

You don’t just get numbers — you understand what’s happening in your system.

New in V3:

  • Node Score → overall health in a single value
  • Load Average (1m / 5m / 15m) → real system load
  • IO Wait → detect disk bottlenecks
  • Node Stress → identify overloaded situations
  • Disk Overload → storage saturation detection
  • Per-core CPU usage → node, VM and container level

:magnifying_glass_tilted_left: Features

  • Full monitoring of:

    • Nodes
    • Virtual Machines (QEMU)
    • Containers (LXC)
    • Storage and disks
    • Proxmox Backup Server (PBS)
  • Advanced infrastructure sensors

  • Maintenance tracking (GC, Prune, Verify)

  • Backup activity and progress

  • Control actions directly from Home Assistant

  • Secure authentication via API tokens

  • Clean and consistent entity structure

  • Optimized updates with low resource usage


:puzzle_piece: PBS Support (V3 Highlight)

V3 introduces full support for Proxmox Backup Server:

  • Datastore usage and status

  • Backup tracking

  • Maintenance operations:

    • Garbage Collection (GC)
    • Prune
    • Verify
  • Deduplication awareness


:warning: Breaking Changes

  • Entity IDs may change due to unique_id fixes
  • A reconfiguration or entity cleanup may be required after updating

:package: Installation

Via HACS (recommended)

  1. Add this repository as a custom repository:
    :backhand_index_pointing_right: GitHub - Javisen/proxmox_sensors: Sensors for Proxmox VE & PBS · GitHub
  2. Install Proxmox Extended Sensors
  3. Restart Home Assistant
  4. Add the integration via UI

:books: Documentation

:backhand_index_pointing_right: GitHub - Javisen/proxmox_sensors: Sensors for Proxmox VE & PBS · GitHub

(Multi-language documentation available)


:speech_balloon: Feedback

Feedback, suggestions, and bug reports are very welcome!

If you try it, I’d love to hear your experience or see how you’re using it in your setup.


Thanks!

I’ve been trying to add this for about an hour.

I created a pve account, gave it admin permissions, and generated a token.

When I go to add the integration, I provide the IP of the management interface, and fill everything out. If I allow automatic node detection, it only detects a single node, and it’s random, so I get duplicates.

If I manually select the node, I am being shown IP addresses that are NOT the management interface - it’s showing one of the node’s cluster IPs, on another one it’s showing the IP for the NFS network…

All that aside, it has never once shown me any of the VMs like the documentation does.

I am using v9.1.6 of PVE. What am I doing wrong?

Thanks for trying the integration :slightly_smiling_face:

From what you describe, this looks more like a configuration/network setup issue rather than a problem with the integration itself — especially in environments with multiple interfaces (cluster, storage, etc.), where Proxmox can return several IPs and it’s not always straightforward to automatically detect the correct one.

Regarding the issues you mentioned:

  • If you’re using a hostname instead of an IP, it may not match the actual node IP.
  • In some setups, the management IP is not attached to a bridge, so cluster or NFS IPs may appear instead.
  • If the wrong node is selected, it’s expected that VMs won’t show up.

I’d recommend trying the following:

  1. Use the node’s management IP directly.
  2. Manually select the correct node.
  3. Verify that the user/token has permissions on /.

That said, I’ll improve the automatic detection in future versions to make it more robust :+1:


If the issue persists, I’d appreciate it if you could open an issue on GitHub with more details (network setup, screenshots, logs, etc.). That makes it easier to track and can help other users facing the same situation.

:point_right: GitHub - Javisen/proxmox_sensors: Sensors for Proxmox VE & PBS · GitHub

1 Like

None of those are the issue, and networking is fine.

I’ll just find a different integration.

No worries :+1:
If you run into the issue again or want to debug it properly, feel free to open an issue on GitHub with details.

Good luck!

1 Like

I’ve already told you what’s wrong, and you all but openly accused me of incompetence rather than look in the mirror. I’m not sure what more you would like from me.

You are the reason this forum is not helpful.

Hello !
Thanks for this integration. Do you mind sharing a nice dashboard ? Like yours in the first post ?
Have a nice day !

I’ve uploaded the sample dashboard to this repository for you to download:

You need to create a folder called “proxmox_sensors” in config/www and include the files from the repository.
Once you rename your sensors, it should look like this:

2 Likes

Hello Javi,

I have 2 problems with your addon, I have 3 proxmox servers with the lm-sensors installed, but none of them shows the temperature on the integration have I missed something?
The other problem is none of the HDDs/NVME/SSD shows smart sensors information

Check if you have the service active with http://YOUR_PROXMOX_IP:9000/sensors

Hello, I’ve followed many lm-sensors guide but none talk about a service http responding on port 9000. It’s an extra service a part from lm-sensors?

Hi!

lm-sensors alone is not enough for this integration.

You also need to install the sensor API service (port 9000), which is explained here:

Quick check:
http://YOUR_PROXMOX_IP:9000/sensors

If that URL does not return JSON data, the service is not installed or not running.

Once that is working, temperatures and SMART data will appear automatically.