Temperatures not reading/being shown in dashboard

[chocolatemalc]

Describe the bug
Temperatures are not being shown in the dashboard.

To Reproduce
Steps to reproduce the behavior:

1. Following the initial setup instructions
2. Fallback method used due to binary not found:

[INFO] Installing temperature monitoring proxy on host...
[INFO] Installing pulse-sensor-proxy for container 200
[INFO] Creating pulse-sensor-proxy service account...
[INFO] Service account created
[INFO] Adding pulse-sensor-proxy to www-data group for Proxmox IPC access...
[INFO] Fetching latest release info...
[INFO] Latest version: v4.23.0
[INFO] Downloading pulse-sensor-proxy-linux-amd64 from GitHub...
curl: (22) The requested URL returned error: 404
[ERROR] Failed to download binary from https://github.com/rcourtman/Pulse/releases/download/v4.23.0/pulse-sensor-proxy-linux-amd64
[WARN] Proxy installation failed - temperature monitoring will use fallback method
[INFO] Check logs: /tmp/proxy-install-200.log
[INFO] Or run manually: curl -fsSL https://raw.githubusercontent.com/rcourtman/Pulse/main/scripts/install-sensor-proxy.sh | bash -s -- --ctid 200
[SUCCESS] Pulse installation complete!

3. I tried running the command manually and the same result.
4. Accessing the web ui and adding my pve node works fine but I've never seen the temperature column.

Expected behavior
I see the temperature column for my PVE node, like I've seen in sample screenshots

Environment:

• Pulse Version: v4.23.0
• Installation Type: ProxmoxVE LXC

Additional context

pulse-diagnostics-sanitized-2025-10-18.json

[rcourtman]

I reproduced the proxy setup log. v4.23.0 release assets do not include the pulse-sensor-proxy binaries so the host script exits and the container falls back to the insecure path, which leaves temperature data unavailable in the UI. I have added the missing binaries to main and they will ship in the next release. I am keeping this open until that build goes out. Once you can update to the upcoming release please rerun curl -fsSL https://raw.githubusercontent.com/rcourtman/Pulse/main/scripts/install-sensor-proxy.sh | bash -s -- --ctid 200 and let me know if the temperature column appears.

[chocolatemalc]

Unsure if you wanted me to wait for the full release.

I rebuilt the CT with v4.24.0-rc.3.
The --version arg in install-sensor-proxy.sh doesn't work because the download logic is ignored unless [[ "$VERSION" == "latest" ]]. I copied the script locally to modify and download the proxy binary from v4.24.0-rc.3

Initial setup went OK as normal

Temp column appears with no reading available:

My sensors output from pve for reference:

~# sensors
k10temp-pci-00c3
Adapter: PCI adapter
Tctl:         +38.8°C  

bnxt_en-pci-0a00
Adapter: PCI adapter
temp1:        +59.0°C  

nvme-pci-0d00
Adapter: PCI adapter
Composite:    +33.9°C  (low  = -273.1°C, high = +80.8°C)
                       (crit = +84.8°C)
Sensor 1:     +36.9°C  (low  = -273.1°C, high = +65261.8°C)
Sensor 2:     +33.9°C  (low  = -273.1°C, high = +65261.8°C)

bnxt_en-pci-0a01
Adapter: PCI adapter
temp1:        +59.0°C  

nvme-pci-0100
Adapter: PCI adapter
Composite:    +32.9°C  (low  = -273.1°C, high = +80.8°C)
                       (crit = +84.8°C)
Sensor 1:     +35.9°C  (low  = -273.1°C, high = +65261.8°C)
Sensor 2:     +32.9°C  (low  = -273.1°C, high = +65261.8°C)

[hamid-elaosta]

I seem to be having the same issue. It appears pulse expects Tdie for AMD CPUs but that's not the case for my server, like @chocolatemalc the correct value to read is Tctl in the k10temp-pci-00c3 device. The JSON looks as follows:

{
  "k10temp-pci-00c3": {
    "Adapter": "PCI adapter",
    "Tctl": {
      "temp1_input": 49.750000
    },
    "Tccd1": {
      "temp3_input": 51.250000
    },
    "Tccd2": {
      "temp4_input": 49.000000
    }
  },
...

This patch to the main branch works for me:

@@ -298,7 +301,7 @@ func (tc *TemperatureCollector) parseCPUTemps(chipMap map[string]interface{}, te
                }
 
                // Look for Package id (Intel) or Tdie (AMD)
-               if strings.Contains(sensorName, "Package id") || strings.Contains(sensorName, "Tdie") {
+               if strings.Contains(sensorName, "Package id") || strings.Contains(sensorName, "Tdie") || strings.Contains(sensorName, "Tctl") {
                        if tempVal := extractTempInput(sensorMap); !math.IsNaN(tempVal) {
                                temp.CPUPackage = tempVal
                                foundPackageTemp = true

[rcourtman]

Appreciate the update. On v4.23.0 the installer looks for pulse-sensor-proxy-linux-amd64 in the release assets but that asset is absent and the tarball also omits bin/pulse-sensor-proxy. Because the host proxy never starts the container remains in the security blocked fallback and temperatures stay empty. v4.25.0 restores the standalone proxy binaries and repackages the tarballs so the installer completes again. After upgrading to v4.25.0 or newer rerun the host installer:

curl -fsSL https://raw.githubusercontent.com/rcourtman/Pulse/main/scripts/install-sensor-proxy.sh | bash -s -- --ctid <container-id>

Then check systemctl status pulse-sensor-proxy, confirm /run/pulse-sensor-proxy/pulse-sensor-proxy.sock exists, and open Diagnostics and click Run Now so the report confirms the socket and daemon status. Temperatures show up again once the proxy responds. Added a reminder on the release checklist to verify these binaries before publishing future tags.

[chocolatemalc]

Will v4.25.0 contain the AMD CPU patch mentioned by @hamid-elaosta ?

[rcourtman]

Will v4.25.0 contain the AMD CPU patch mentioned by @hamid-elaosta ?

No sorry that fix came after 4.25, but will be in next release. You can test it on your host by running curl -fsSL https://raw.githubusercontent.com/rcourtman/Pulse/main/install.sh | bash -s -- --source.

[chocolatemalc]

No problem.

Build failed:

=================================================
           Pulse Installation Script
=================================================

[INFO] Installing dependencies...
[INFO] Creating pulse user...
[INFO] Setting up directories...
[INFO] Building Pulse from source (branch: main)...
[INFO] Installing build dependencies...
[ERROR] Failed to build backend
[ERROR] Source build failed

However, I'm not in a rush, I can wait for the next stable release. I really appreciate your work.

[syselement]

Appreciate the update. On v4.23.0 the installer looks for pulse-sensor-proxy-linux-amd64 in the release assets but that asset is absent and the tarball also omits bin/pulse-sensor-proxy. Because the host proxy never starts the container remains in the security blocked fallback and temperatures stay empty. v4.25.0 restores the standalone proxy binaries and repackages the tarballs so the installer completes again. After upgrading to v4.25.0 or newer rerun the host installer:

curl -fsSL https://raw.githubusercontent.com/rcourtman/Pulse/main/scripts/install-sensor-proxy.sh | bash -s -- --ctid <container-id>

Then check systemctl status pulse-sensor-proxy, confirm /run/pulse-sensor-proxy/pulse-sensor-proxy.sock exists, and open Diagnostics and click Run Now so the report confirms the socket and daemon status. Temperatures show up again once the proxy responds. Added a reminder on the release checklist to verify these binaries before publishing future tags.

hey!

Environment:

• Pulse Version: v4.25.0
• Installation Type: ProxmoxVE LXC

I've tried the

curl -fsSL https://raw.githubusercontent.com/rcourtman/Pulse/main/scripts/install-sensor-proxy.sh | bash -s -- --ctid <container-id>

command by debugging it with:

export PS4='+ ${BASH_SOURCE}:${LINENO}:${FUNCNAME[0]} '
bash -euxo pipefail -x /tmp/install-sensor-proxy.sh --ctid <container-id>

finding out it was stopping at:

[...]
[INFO] Proxy public key: ssh-ed25519 AAAAC3Nza...
+ command -v pvecm
++ pvecm status
++ awk '/0x[0-9a-f]+.*[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+/ {print $3}'
+ CLUSTER_NODES=
root@pve:~# 

I think on single-node Proxmox installations (no cluster), the installer exits immediately after computing CLUSTER_NODES because the pipeline

CLUSTER_NODES=$(pvecm status 2>/dev/null | awk '...')

runs under set -euo pipefail, and pvecm status returns a non-zero exit code when no cluster is configured.

• With set -o pipefail, the pipeline inherits that non-zero status and the script exits before reaching the “standalone node” else branch.
• With set -e, the script exits on that non-zero status, skipping the subsequent if [[ -n "$CLUSTER_NODES" ]] … else … fi.

# e.g. on PVE host
root@pve:~# pvecm status
Error: Corosync config '/etc/pve/corosync.conf' does not exist - is this node part of a cluster?
root@pve:~# pvecm status >/dev/null 2>&1; echo $?
2

Expected Behavior

On a standalone node, the script should continue and run the else branch:

• “No cluster detected…”
• “Authorizing proxy key on localhost…”
• etc

Working fix (the only one that I've tried)

At line 979, make the probe tolerant by forcing success on pvecm status and letting the awk decide if any nodes exist:

CLUSTER_NODES=$(
  ( pvecm status 2>/dev/null || true ) | awk '/0x[0-9a-f]+.*[0-9]+\.[0-9]+\.[0-9]+/ {print $3}'
)

# e.g. on PVE host
root@pve:~# pvecm status >/dev/null 2>&1 || true ; echo $?
0

• On clustered nodes, pvecm status exits 0 -> behavior unchanged
• On standalone nodes, pvecm status exits non-zero but is neutralized by || true, allowing awk to read an empty stream and set CLUSTER_NODES=""
• The subsequent if [[ -n "$CLUSTER_NODES" ]] correctly falls into the standalone else path, finishes all the remaining install-sensor-proxy.sh script parts correctly and temperature is showing in the GUI after the Pulse LXC restart

Some other fixes may be interesting, but I'm not a bash expert.

PS:
This may be necessary on 804 too but not sure.

[BunpGhost]

Hi
I think I have the same issue.
Just did a clean install.
Then installed the proxy on host.
And no temperatures show up.
I don't know how to debug. THe only thing I did was to get the status of the proxy service. I found this error:

Oct 31 15:36:15 pve pulse-sensor-proxy[2066501]: 3:36PM WRN Failed to get temperatures error="failed to fetch temperatures: exit status 255" corr_id=70240b58-ce7d-4efc-9a01-9039917eb4f1 method=get_temperature node=192.168.0.126 pid=2056198 uid=100999
Oct 31 15:36:15 pve pulse-sensor-proxy[2066501]: 3:36PM WRN Handler failed error="failed to get temperatures: failed to fetch temperatures: exit status 255" corr_id=70240b58-ce7d-4efc-9a01-9039917eb4f1 method=get_temperature pid=2056198 uid=100999
Oct 31 15:36:15 pve pulse-sensor-proxy[2066501]: 3:36PM DBG Peer credentials gid=100996 pid=2056198 uid=100999
Oct 31 15:36:15 pve pulse-sensor-proxy[2066501]: 3:36PM DBG Fetching temperature via SSH corr_id=eef821ba-1ee6-472d-9c15-57d190a5b9f3 method=get_temperature node=192.168.0.126 pid=2056198 uid=100999

Any help?

[chocolatemalc]

Temperatures were working for a few hours after upgrading to 4.26. Stopped working after update to 4.26.1. Haven't been able to get it working since.

Latest fresh LXC attempt today:

============================================
  Pulse Monitoring Setup for Proxmox VE
============================================

Detected Proxmox VE host environment.


What would you like to do?

  [1] Install/Configure - Set up Pulse monitoring
  [2] Remove All        - Uninstall everything Pulse has configured
  [3] Cancel            - Exit without changes

Your choice [1/2/3]: 1

Proceeding with installation...

Checking for existing Pulse monitoring tokens from this Pulse server...
Creating monitoring user...
Generating API token...
API token generated successfully

Registering node with Pulse...
Node registered successfully


Setting up permissions...
Setting up additional permissions...
  • Applied privileges: Sys.Audit VM.GuestAgent.Audit

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Temperature Monitoring Setup (Optional)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📊 Enable Temperature Monitoring?

Collect CPU and drive temperatures via secure SSH connection.

Security:
  • SSH key authentication with forced command (sensors -j only)
  • No shell access, port forwarding, or other SSH features
  • Keys stored in Pulse service user's home directory

Enable temperature monitoring? [y/N]
> y

Configuring temperature monitoring...
  ✓ Sensors key configured (restricted to sensors -j)
  ✓ lm-sensors package verified

✓ Temperature monitoring enabled
  Temperature data will appear in the dashboard within 10 seconds

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Standalone Node Temperature Setup
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Detected: This is a standalone node (not in a Proxmox cluster)

For enhanced security with containerized Pulse, we'll fetch the
temperature proxy's SSH key directly from your Pulse server.

Fetching temperature proxy public key...
  ✓ Retrieved proxy public key
  ✓ Proxy key already configured

✓ Standalone node temperature monitoring configured
  The Pulse temperature proxy can now collect temperature data
  from this node using secure SSH with forced commands.


━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Setup Complete
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Node successfully registered with Pulse monitoring.
Data will appear in your dashboard within 10 seconds.

Something going on with the proxy:

pulse-diagnostics-sanitized-2025-11-09.json

 pulse-sensor-proxy.service - Pulse Temperature Proxy
     Loaded: loaded (/etc/systemd/system/pulse-sensor-proxy.service; enabled; preset: enabled)
     Active: active (running) since Mon 2025-11-10 00:02:47 AEDT; 21min ago
 Invocation: 0b29aa0a14f14cce921f5f5139d403ab
       Docs: https://github.com/rcourtman/Pulse
   Main PID: 2387679 (pulse-sensor-pr)
      Tasks: 13 (limit: 74014)
     Memory: 8.6M (peak: 56.4M)
        CPU: 9.598s
     CGroup: /system.slice/pulse-sensor-proxy.service
             └─2387679 /usr/local/bin/pulse-sensor-proxy

Nov 10 00:23:34 pve pulse-sensor-proxy[2387679]: 12:23AM WRN Handler failed error="failed to evaluate cluster membership: failed to get cluster status: exi>
Nov 10 00:23:35 pve pulse-sensor-proxy[2387679]: 12:23AM WRN pvecm status failed stderr="ipcc_send_rec[1] failed: Unknown error -1\nipcc_send_rec[2] failed>
Nov 10 00:23:35 pve pulse-sensor-proxy[2387679]: 12:23AM WRN Node validation failed error="failed to evaluate cluster membership: failed to get cluster sta>
Nov 10 00:23:35 pve pulse-sensor-proxy[2387679]: 12:23AM WRN Handler failed error="failed to evaluate cluster membership: failed to get cluster status: exi>
Nov 10 00:23:35 pve pulse-sensor-proxy[2387679]: 12:23AM WRN pvecm status failed stderr="ipcc_send_rec[1] failed: Unknown error -1\nipcc_send_rec[2] failed>
Nov 10 00:23:35 pve pulse-sensor-proxy[2387679]: 12:23AM WRN Node validation failed error="failed to evaluate cluster membership: failed to get cluster sta>
Nov 10 00:23:35 pve pulse-sensor-proxy[2387679]: 12:23AM WRN Handler failed error="failed to evaluate cluster membership: failed to get cluster status: exi>
Nov 10 00:23:36 pve pulse-sensor-proxy[2387679]: 12:23AM WRN pvecm status failed stderr="ipcc_send_rec[1] failed: Unknown error -1\nipcc_send_rec[2] failed>
Nov 10 00:23:36 pve pulse-sensor-proxy[2387679]: 12:23AM WRN Node validation failed error="failed to evaluate cluster membership: failed to get cluster sta>
Nov 10 00:23:36 pve pulse-sensor-proxy[2387679]: 12:23AM WRN Handler failed error="failed to evaluate cluster membership: failed to get cluster status: exi>
~

[chocolatemalc]

Attempted to debug service on PVE further:

~#pulse-sensor-proxy
12:27AM WRN No allowed_source_subnets configured; using detected host addresses (recommended to configure explicitly) auto_detected_subnets=["xxx.xxx.xxx.xxx/32"]
12:27AM FTL Failed to drop privileges error="lookup user \"pulse-sensor\" failed: user: unknown user pulse-sensor (fallback: user \"pulse-sensor\" not found in /etc/passwd)" user=pulse-sensor

~# adduser --system --no-create-home --group pulse-sensor

~# pulse-sensor-proxy
12:30AM WRN No allowed_source_subnets configured; using detected host addresses (recommended to configure explicitly) auto_detected_subnets=["xxx.xxx.xxx.xxx/32"]
12:30AM INF Running as unprivileged user gid=112 uid=110 user=pulse-sensor
12:30AM WRN Cannot open audit log file, falling back to stderr (systemd journal) error="open /var/log/pulse/sensor-proxy/audit.log: permission denied" path=/var/log/pulse/sensor-proxy/audit.log
12:30AM WRN Audit logger initialized with stderr fallback due to filesystem constraints mode=stderr path=/var/log/pulse/sensor-proxy/audit.log
12:30AM INF Node validator using Proxmox cluster membership (auto-detect)
12:30AM INF Starting pulse-sensor-proxy audit_log=/var/log/pulse/sensor-proxy/audit.log config_path=/etc/pulse-sensor-proxy/config.yaml log_level=info socket=/run/pulse-sensor-proxy/pulse-sensor-proxy.sock ssh_key_dir=/var/lib/pulse-sensor-proxy/ssh version=v4.27.1
12:30AM WRN Failed to determine working directory; using default error="stat .: permission denied"
12:30AM INF Loaded subordinate ID ranges for ID-mapped root authentication gid_range_count=2 uid_range_count=2 users=["root"]
12:30AM FTL Failed to start proxy error="failed to create SSH key directory: mkdir /var/lib/pulse-sensor-proxy/ssh: permission denied"

~# usermod -aG pulse-sensor-proxy pulse-sensor
~# chmod -R 770 /var/lib/pulse-sensor-proxy
~# chmod 770 /var/log/pulse

~# pulse-sensor-proxy
12:41AM WRN No allowed_source_subnets configured; using detected host addresses (recommended to configure explicitly) auto_detected_subnets=["xxx.xxx.xxx.xxx/32"]
12:41AM INF Running as unprivileged user gid=112 uid=110 user=pulse-sensor
12:41AM WRN Cannot open audit log file, falling back to stderr (systemd journal) error="open /var/log/pulse/sensor-proxy/audit.log: permission denied" path=/var/log/pulse/sensor-proxy/audit.log
12:41AM WRN Audit logger initialized with stderr fallback due to filesystem constraints mode=stderr path=/var/log/pulse/sensor-proxy/audit.log
12:41AM INF Node validator using Proxmox cluster membership (auto-detect)
12:41AM INF Starting pulse-sensor-proxy audit_log=/var/log/pulse/sensor-proxy/audit.log config_path=/etc/pulse-sensor-proxy/config.yaml log_level=info socket=/run/pulse-sensor-proxy/pulse-sensor-proxy.sock ssh_key_dir=/var/lib/pulse-sensor-proxy/ssh version=v4.27.1
12:41AM WRN Failed to determine working directory; using default error="stat .: permission denied"
12:41AM INF Loaded subordinate ID ranges for ID-mapped root authentication gid_range_count=2 uid_range_count=2 users=["root"]
12:41AM INF SSH keypair already exists
12:41AM INF Unix socket ready socket=/run/pulse-sensor-proxy/pulse-sensor-proxy.sock
12:41AM INF Metrics server started addr=127.0.0.1:9127
^C12:41AM INF Shutting down proxy...
12:41AM INF Proxy stopped

Still no temps shown.