Hi, I just found this integration and it implements all the missing pieces I want.
But I am facing an issue when I want to add the individual nodes.
I already added PBS and the Cluster, but every time I want to add a node, it provides the name of total different node, for example proxmox01 adds proxmox03.
Any idea what this could be?
[edit] already fixed the issue for the vm/lxc, the api token also needed permissions it seems. [/edit]
Is it better to create an issue on github?
Thanks for the report.
A couple of questions:
Which permissions does your API token have? Is it the same token you use for both the Cluster and the individual PVE entries?
Are you connecting directly to each node (e.g. proxmox01:8006) or through a shared hostname/VIP?
Which Proxmox VE version are you running?
The "Unknown" cluster sensors and missing VMs/CTs are often caused by insufficient API permissions, so I'd like to rule that out first.
If possible, could you please open a GitHub issue instead?
The issue template collects important information such as your Home Assistant version, Proxmox VE version, integration version, installation method, logs, and other details that will help me investigate the problem much more efficiently.
Since I tested the installation flow from scratch before releasing v4.0.1 and couldn't reproduce this behavior, having all that information in one place will make it much easier to determine whether this is an environment-specific issue or a bug in the integration.
Thanks!
Did so, but haven’t seen a reaction yet in the maintime 
@Javisendra seems to have abandoned us. I'm uninstalling and going to Glances instead.
Perhaps on holiday or in hospitaal? Don't judge what you don't know for sure.
Thanks for reporting the issues. I've been away for a few days and I'm catching up now. I'll review the open issues and start addressing them.
It's fixed and tested now. This was my mistake: during a previous merge, an older version of the config_flow.py file accidentally slipped back into the project. Fortunately, the issue was easy to identify and fix.
Sorry for the delay, and thanks for your patience. I'll move on to the next issue now.
Since the integration is officially included in HACS, the update should be available shortly. If you don't want to wait, you can open HACS, go to the integration page, and the new release should already be available.
v4.0.2 – Fix Config Entry version regression
Hi,
Thank you for the detailed report. After reviewing the issue, I believe I’ve identified the root cause.
The problem is related to the automatic node detection logic. When the integration cannot reliably match the configured host/IP to a specific Proxmox node, it currently falls back to selecting the first node returned by the cluster API. In environments using a VIP, load balancer, or certain network configurations, this can result in the wrong node being selected.
I’m going to change this behavior. Instead of silently falling back to the first node, the integration will notify the user that automatic detection failed and allow manual selection from the available cluster nodes. This should make the behavior much more predictable and avoid incorrect node assignments.
This change affects the configuration flow, so I want to implement it carefully and test it thoroughly before releasing a fix. It may therefore take a little longer than a simple bug fix, but I’d rather ensure the solution is robust than introduce new issues.
Thanks for your patience. I’m actively working on it and will update this issue as soon as I have the patch ready.
Hey Javisendra,
I really enjoy using your integration!
Where do you want feature requests to be placed?
I would like to request the following:
Just recently I was looking for a way to tell which node a VM/LXC is connected to. This is not possible currently.
So the function via_device might be worth looking at to be used in your integration.
However frenck did start a discussion about an alternative way just a few weeks ago: Sub-devices: a first-class parent/child device relationship · home-assistant/architecture · Discussion #1414 · GitHub
Maybe it’s worth a read or even chiming in 
I like this integration that can give me a lot more info. However, the Voltage Rails reading was reported as “Temperature”. Will be good if these can be corrected.
Thanks for reporting this.
Could you please open the following URL in your browser?
http://<your-proxmox-ip>:9000/sensors
and send me the JSON output (or at least the section containing the It8772 sensor)?
I don’t have a motherboard that exposes voltage sensors, so I’d like to see how they’re reported by the API in order to correctly identify them and assign the proper Home Assistant unit (V instead of °C).
This is the output from “:9000/sensors”
{“soc_dts1-virtual-0”:{“Adapter”:“Virtual device”,“temp1”:{“temp1_input”:44.000000}},“coretemp-isa-0000”:{“Adapter”:“ISA adapter”,“Core 0”:{“temp2_input”:41.000000,“temp2_max”:105.000000,“temp2_crit”:105.000000,“temp2_crit_alarm”:0.000000},“Core 1”:{“temp3_input”:41.000000,“temp3_max”:105.000000,“temp3_crit”:105.000000,“temp3_crit_alarm”:0.000000},“Core 2”:{“temp4_input”:45.000000,“temp4_max”:105.000000,“temp4_crit”:105.000000,“temp4_crit_alarm”:0.000000},“Core 3”:{“temp5_input”:45.000000,“temp5_max”:105.000000,“temp5_crit”:105.000000,“temp5_crit_alarm”:0.000000}},“it8772-isa-0a40”:{“Adapter”:“ISA adapter”,“in0”:{“in0_input”:1.632000,“in0_min”:2.268000,“in0_max”:3.060000,“in0_alarm”:1.000000,“in0_beep”:0.000000},“in1”:{“in1_input”:1.356000,“in1_min”:2.676000,“in1_max”:2.100000,“in1_alarm”:1.000000,“in1_beep”:0.000000},“in2”:{“in2_input”:2.208000,“in2_min”:0.204000,“in2_max”:2.064000,“in2_alarm”:1.000000,“in2_beep”:0.000000},“+3.3V”:{“in3_input”:3.288000,“in3_min”:2.256000,“in3_max”:5.232000,“in3_alarm”:0.000000,“in3_beep”:0.000000},“in4”:{“in4_input”:2.232000,“in4_min”:2.580000,“in4_max”:1.620000,“in4_alarm”:1.000000,“in4_beep”:0.000000},“in5”:{“in5_input”:2.232000,“in5_min”:1.836000,“in5_max”:0.132000,“in5_alarm”:1.000000,“in5_beep”:0.000000},“in6”:{“in6_input”:2.220000,“in6_min”:1.524000,“in6_max”:2.292000,“in6_alarm”:0.000000,“in6_beep”:0.000000},“3VSB”:{“in7_input”:3.264000,“in7_min”:5.352000,“in7_max”:1.440000,“in7_alarm”:1.000000,“in7_beep”:0.000000},“Vbat”:{“in8_input”:3.312000},“fan1”:{“fan1_input”:0.000000,“fan1_min”:0.000000,“fan1_alarm”:0.000000,“fan1_beep”:0.000000},“fan2”:{“fan2_input”:0.000000,“fan2_min”:0.000000,“fan2_alarm”:0.000000,“fan2_beep”:0.000000},“fan3”:{“fan3_input”:0.000000,“fan3_min”:0.000000,“fan3_alarm”:0.000000,“fan3_beep”:0.000000},“temp1”:{“temp1_input”:-38.000000,“temp1_max”:127.000000,“temp1_min”:-128.000000,“temp1_alarm”:0.000000,“temp1_type”:6.000000,“temp1_offset”:86.000000,“temp1_beep”:0.000000},“temp2”:{“temp2_input”:38.000000,“temp2_max”:127.000000,“temp2_min”:-128.000000,“temp2_alarm”:0.000000,“temp2_type”:3.000000,“temp2_offset”:25.000000,“temp2_beep”:0.000000},“temp3”:{“temp3_input”:-56.000000,“temp3_max”:127.000000,“temp3_min”:48.000000,“temp3_alarm”:0.000000,“temp3_type”:3.000000,“temp3_offset”:0.000000,“temp3_beep”:0.000000},“pwm1”:{“pwm1”:24.000000,“pwm1_freq”:23437.000000,“pwm1_enable”:2.000000},“pwm2”:{“pwm2”:24.000000,“pwm2_freq”:23437.000000,“pwm2_enable”:2.000000},“pwm3”:{“pwm3”:64.000000,“pwm3_freq”:23437.000000,“pwm3_enable”:0.000000},“intrusion0”:{“intrusion0_alarm”:1.000000}},“soc_dts0-virtual-0”:{“Adapter”:“Virtual device”,“temp1”:{“temp1_input”:43.000000}}}
[Feature Request]
One of my nodes is powered up only for a few minutes every day, so that it can sync my backups.
Ofc I want to monitor that node via HA too.
However the log get’s spammed with these messages:
2026-07-12 08:01:18.739 ERROR (MainThread) [custom_components.proxmox_sensors.api] PVE GET error on nodes/node3/status: HTTPSConnectionPool(host='node3.domain.tld', port=8006): Max retries exceeded with url: /api2/json/nodes/node3/status (Caused by NewConnectionError("HTTPSConnection(host='node3.domain.tld', port=8006): Failed to establish a new connection: [Errno 113] Host is unreachable"))
I would like to request to suppress messages like that. Maybe with an option per node “check online status prior querying” or something similar.
Upcoming update
I’ve been working on the reported issues and all three have now been addressed. These fixes will be included in the next release.
lm-sensors are now exposed with the correct Voltage (V) unit instead of being incorrectly shown as temperatures.lm-sensors.Thanks to everyone who reported these issues and provided logs and testing. Your feedback helped improve the integration for different Proxmox environments.
NEW RELEASE 4.0.3
Thanks to everyone who reported issues, provided logs, and helped test these improvements.