Temperature monitoring with Pulse deployed in Kubernetes

[zkutasi]

Is your feature request related to a problem? Please describe.
I have observed, that for sensor/temp data, I would need to have a pulse-sensor-proxy... here: https://github.com/rcourtman/Pulse/blob/main/docs/TEMPERATURE_MONITORING.md I can see that there are ways to install in LXC, Docker and remote... but it does not mention how to install this when in Kubernetes.

My setup is, that Pulse runs in Kubernetes, and I have two Proxmox systems and 4-5 Docker hosts.

I do not know where this proxy shall be at all: should it be a sidecar in the Pod? shall it be installed remotely? I think as you mention that in case of docker, one must install this on the host, it should be a sidecar.

Describe the solution you'd like
Explain how to set up temp monitoring for Kubernetes-centralized Pulse, or explain why it is not supported. If not supported, maybe also not allow to be enabled at all.

Describe alternatives you've considered
N/A

Additional context
N/A

[rcourtman]

Temperature monitoring requires access to hardware sensors which are only available from the physical host - containers (whether Docker or K8s) can't access them.

For your setup with Pulse in Kubernetes:

For Proxmox nodes: Install the sensor-proxy directly on each Proxmox host using the "remote" installation method from Settings → Host Agents. This runs outside Pulse's container and sends temperature data back.

For Docker hosts: Same approach - install sensor-proxy on the bare metal or VM running Docker, not inside a container.

A sidecar wouldn't work because it would still be containerized and unable to access host sensors. The proxy needs to run natively on each host you want temperature data from.

The install commands from Settings → Host Agents should work for both cases - just run them on each physical host rather than inside containers.

[zkutasi]

OK, now I understand.
So I have executed the sensor install on my Proxmox cluster Node1:

root@proxmox1:~# curl -fsSL https://github.com/rcourtman/Pulse/releases/latest/download/install-sensor-proxy.sh | sudo bash -s -- --standalone --http-mode --pulse-server http://pulse.kubernetes.home
[INFO] Installing pulse-sensor-proxy for standalone/Docker deployment
[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] Setting up installation directories...
[WARN] Failed to resolve latest GitHub release (HTTP 22)
[INFO] Downloading pulse-sensor-proxy-linux-amd64 from fallback source...
[INFO] Binary installed to /opt/pulse/sensor-proxy/bin/pulse-sensor-proxy
[INFO] Setting up service directories with proper ownership...
[INFO] Detected pulse-sensor-proxy v4.33.1 (allowed_nodes_file supported)
[INFO] Creating base configuration file...

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  HTTP Mode Setup (External PVE Host)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

[INFO] Testing connection to Pulse server...
✓ Pulse server is reachable
[INFO] Generating self-signed TLS certificate...
✓ TLS certificate generated (SHA256: AE:59:E1:73:95:CA:01:90:B9:AD:54:03:A2:B4:31:38:AB:EE:87:8A:A6:D8:2B:F9:43:ED:DD:18:89:3F:DC:23)
[INFO] Proxy will be accessible at: https://192.168.31.11:8443
[INFO] Registering temperature proxy with Pulse at http://pulse.kubernetes.home...
✓ Registered successfully
[INFO] Config backed up to: /etc/pulse-sensor-proxy/config.yaml.backup.1764779131
[INFO] Pulse server detected at: 192.168.31.31
[INFO] Configuring HTTP mode...
[INFO] Added allowed_source_subnets entry 192.168.31.31/32
[INFO] Added allowed_source_subnets entry 127.0.0.1/32
[INFO] Added allowed_source_subnets entry 192.168.31.11/32
[INFO] HTTP mode configured successfully (using binary)
✓ HTTP mode configured successfully

[INFO] Firewall configuration required:
[INFO]   Allow inbound TCP connections on port 8443 from Pulse server
[INFO]   Command: ufw allow from <pulse-server-ip> to any port 8443

[INFO] Writing control plane token...
[INFO] Installing hardened systemd service...
[INFO] Updating service file for HTTP mode...
✓ Service file updated for HTTP mode
[INFO] Enabling and starting service...
Created symlink /etc/systemd/system/multi-user.target.wants/pulse-sensor-proxy.service → /etc/systemd/system/pulse-sensor-proxy.service.
[INFO] Waiting for socket...
[INFO] Socket ready at /run/pulse-sensor-proxy/pulse-sensor-proxy.sock
[INFO] Validating HTTP endpoint...
✓ HTTP endpoint validated successfully
[INFO] Installing sensor wrapper script...
✓ Sensor wrapper installed at /opt/pulse/sensor-proxy/bin/pulse-sensor-wrapper.sh
[INFO] Installing cleanup system...
[INFO] Cleanup script installed
Created symlink /etc/systemd/system/multi-user.target.wants/pulse-sensor-cleanup.path → /etc/systemd/system/pulse-sensor-cleanup.path.
[INFO] Cleanup system installed and enabled
[INFO] Configuring proxy SSH access to cluster nodes...
[INFO] Proxy public key: ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIO3cVMvRwLsI+1...

Then I had to extend the allowed_source_subnets, to 192.168.31.0/24.

/opt/pulse/sensor-proxy/bin/pulse-sensor-proxy config add-subnet 192.168.31.0/24

But still I cannot seem to get this to work.

For a brief moment when I restart the proxy service, I see a temperature, then nothing.
I also get sometimes connection issues on the main UI for the WebSocket... and after a few seconds, it is back again. Not sure if it my ingress (Contour) or something else.

[rcourtman]

The screenshot shows HTTPS timeouts connecting to the temperature proxy at 192.168.31.11:8443. The "context deadline exceeded" errors mean Pulse can't reach the proxy endpoint at all.

A few things to check:

1. Verify the proxy is listening on the correct interface:

   # On proxmox1
   ss -tlnp | grep 8443

   Should show something like 0.0.0.0:8443 or 192.168.31.11:8443

2. Test connectivity from your Kubernetes cluster:

   # From a pod in the same namespace as Pulse
   curl -k https://192.168.31.11:8443/health

3. Check firewall rules on Proxmox:

   # On proxmox1
   iptables -L -n | grep 8443
   # Or if using pve-firewall
   pve-firewall status

The brief temperature appearance when restarting the proxy might be a timing issue - the proxy responds to the first request but then subsequent requests timeout.

Could you share the output of those commands? Also check journalctl -u pulse-sensor-proxy -n 50 for any errors when Pulse tries to connect.

[zkutasi]

root@proxmox1:~# ss -tlnp | grep 8443
LISTEN 0      128                *:8443             *:*    users:(("pulse-sensor-pr",pid=3288661,fd=8))

Weird, even localhost access is not working.

root@proxmox1:~# timeout 15 curl -k https://127.0.0.1:8443/health
root@proxmox1:~#

Nothing in there. I have disabled the pve-firewall now though, lets see.

Here are the last 50 loglines:

root@proxmox1:~# journalctl -u pulse-sensor-proxy.service | tail -50
Dec 03 19:41:00 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox3: knownhosts: ssh-keyscan failed for proxmox3:22: exit status 1 (output: getaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known)" node=proxmox3
Dec 03 19:41:00 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox2: knownhosts: ssh-keyscan failed for proxmox2:22: exit status 1 (output: getaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known)" node=proxmox2
Dec 03 19:41:00 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM WRN SSH temperature collection failed on local node, falling back to direct sensors error="exit status 255" node=proxmox1
Dec 03 19:41:00 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM INF Temperature data fetched successfully via HTTP node=proxmox1
Dec 03 19:41:20 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM INF Updated node allow-list from control plane allowed_cidr_count=0 allowed_node_count=7
Dec 03 19:41:30 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox1
Dec 03 19:41:30 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox3
Dec 03 19:41:30 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox3: knownhosts: ssh-keyscan failed for proxmox3:22: exit status 1 (output: getaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known)" node=proxmox3
Dec 03 19:41:30 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox2
Dec 03 19:41:30 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox2: knownhosts: ssh-keyscan failed for proxmox2:22: exit status 1 (output: getaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known)" node=proxmox2
Dec 03 19:41:30 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM WRN SSH temperature collection failed on local node, falling back to direct sensors error="exit status 255" node=proxmox1
Dec 03 19:41:31 proxmox1 pulse-sensor-proxy[3288661]: 7:41PM INF Temperature data fetched successfully via HTTP node=proxmox1
Dec 03 19:42:01 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox3
Dec 03 19:42:01 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox2
Dec 03 19:42:01 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox3: knownhosts: ssh-keyscan failed for proxmox3:22: exit status 1 (output: getaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known)" node=proxmox3
Dec 03 19:42:01 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox1
Dec 03 19:42:01 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox2: knownhosts: ssh-keyscan failed for proxmox2:22: exit status 1 (output: getaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known)" node=proxmox2
Dec 03 19:42:01 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN SSH temperature collection failed on local node, falling back to direct sensors error="exit status 255" node=proxmox1
Dec 03 19:42:01 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM INF Temperature data fetched successfully via HTTP node=proxmox1
Dec 03 19:42:20 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM INF Updated node allow-list from control plane allowed_cidr_count=0 allowed_node_count=7
Dec 03 19:42:50 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox3
Dec 03 19:42:50 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox2
Dec 03 19:42:50 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox1
Dec 03 19:42:50 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox3: knownhosts: ssh-keyscan failed for proxmox3:22: exit status 1 (output: getaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known)" node=proxmox3
Dec 03 19:42:50 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox2: knownhosts: ssh-keyscan failed for proxmox2:22: exit status 1 (output: getaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known)" node=proxmox2
Dec 03 19:42:50 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM WRN SSH temperature collection failed on local node, falling back to direct sensors error="exit status 255" node=proxmox1
Dec 03 19:42:50 proxmox1 pulse-sensor-proxy[3288661]: 7:42PM INF Temperature data fetched successfully via HTTP node=proxmox1
Dec 03 19:43:20 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM INF Updated node allow-list from control plane allowed_cidr_count=0 allowed_node_count=7
Dec 03 19:43:20 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox1
Dec 03 19:43:20 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox3
Dec 03 19:43:20 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox2
Dec 03 19:43:20 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox2: knownhosts: ssh-keyscan failed for proxmox2:22: exit status 1 (output: getaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known)" node=proxmox2
Dec 03 19:43:20 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox3: knownhosts: ssh-keyscan failed for proxmox3:22: exit status 1 (output: getaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known)" node=proxmox3
Dec 03 19:43:20 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN SSH temperature collection failed on local node, falling back to direct sensors error="exit status 255" node=proxmox1
Dec 03 19:43:20 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM INF Temperature data fetched successfully via HTTP node=proxmox1
Dec 03 19:43:50 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox3
Dec 03 19:43:50 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox2
Dec 03 19:43:50 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox3: knownhosts: ssh-keyscan failed for proxmox3:22: exit status 1 (output: getaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known)" node=proxmox3
Dec 03 19:43:50 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox1
Dec 03 19:43:50 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox2: knownhosts: ssh-keyscan failed for proxmox2:22: exit status 1 (output: getaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known)" node=proxmox2
Dec 03 19:43:51 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM WRN SSH temperature collection failed on local node, falling back to direct sensors error="exit status 255" node=proxmox1
Dec 03 19:43:51 proxmox1 pulse-sensor-proxy[3288661]: 7:43PM INF Temperature data fetched successfully via HTTP node=proxmox1
Dec 03 19:44:20 proxmox1 pulse-sensor-proxy[3288661]: 7:44PM INF Updated node allow-list from control plane allowed_cidr_count=0 allowed_node_count=7
Dec 03 19:44:40 proxmox1 pulse-sensor-proxy[3288661]: 7:44PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox3
Dec 03 19:44:40 proxmox1 pulse-sensor-proxy[3288661]: 7:44PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox1
Dec 03 19:44:40 proxmox1 pulse-sensor-proxy[3288661]: 7:44PM WRN Failed to load host key from Proxmox; falling back to ssh-keyscan error="open /etc/pve/priv/known_hosts: permission denied" node=proxmox2
Dec 03 19:44:40 proxmox1 pulse-sensor-proxy[3288661]: 7:44PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox3: knownhosts: ssh-keyscan failed for proxmox3:22: exit status 1 (output: getaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known\r\ngetaddrinfo proxmox3: Name or service not known)" node=proxmox3
Dec 03 19:44:40 proxmox1 pulse-sensor-proxy[3288661]: 7:44PM WRN Failed to get temperatures via SSH error="failed to ensure host key for proxmox2: knownhosts: ssh-keyscan failed for proxmox2:22: exit status 1 (output: getaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known\r\ngetaddrinfo proxmox2: Name or service not known)" node=proxmox2
Dec 03 19:44:40 proxmox1 pulse-sensor-proxy[3288661]: 7:44PM WRN SSH temperature collection failed on local node, falling back to direct sensors error="exit status 255" node=proxmox1
Dec 03 19:44:40 proxmox1 pulse-sensor-proxy[3288661]: 7:44PM INF Temperature data fetched successfully via HTTP node=proxmox1
root@proxmox1:~#

[zkutasi]

Hmmmm, it seems it is bound to only IPv6?

root@proxmox1:~# curl -k https://[::1]:8443
{"error":"source IP not allowed"}
root@proxmox1:~#

IPv4 localhost is hanging. This is a Proxmox v8 host.

I have added forcefully the http-address to bind to:

curl -fsSL https://github.com/rcourtman/Pulse/releases/latest/download/install-sensor-proxy.sh | sudo bash -s -- --standalone --http-mode --pulse-server http://pulse.kubernetes.home --http-addr 0.0.0.0:8443

Mind you... now the github download also works, so my Proxmox prefers IPv6 then?

I think also there is a small typo in the install script, in the testing part:

root@proxmox1:~# curl -fsSL https://github.com/rcourtman/Pulse/releases/latest/download/install-sensor-proxy.sh | sudo bash -s -- --standalone --http-mode --pulse-server http://pulse.kubernetes.home --http-addr 0.0.0.0:8443
[INFO] Installing pulse-sensor-proxy for standalone/Docker deployment
[INFO] Setting up installation directories...
[INFO] Downloading pulse-sensor-proxy-linux-amd64 from GitHub release v4.36.1...
[INFO] Downloading pulse-v4.36.1-linux-amd64.tar.gz to extract pulse-sensor-proxy...
[INFO] Binary installed to /opt/pulse/sensor-proxy/bin/pulse-sensor-proxy
[INFO] Setting up service directories with proper ownership...
[INFO] Detected pulse-sensor-proxy v4.36.1 (allowed_nodes_file supported)
Migration complete: inline allowed_nodes moved to file
✓ Migration complete: inline allowed_nodes moved to file

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  HTTP Mode Setup (External PVE Host)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

[INFO] Testing connection to Pulse server...
✓ Pulse server is reachable
[INFO] Generating self-signed TLS certificate...
✓ TLS certificate generated (SHA256: 85:E8:FF:B4:DA:FD:DD:59:81:D8:3A:DE:0C:F9:D3:66:24:8C:FA:BC:CB:92:4B:A1:DD:BE:F9:27:E1:12:7F:3F)
[INFO] Proxy will be accessible at: https://192.168.31.110.0.0.0:8443
[INFO] Registering temperature proxy with Pulse at http://pulse.kubernetes.home...
✓ Registered successfully
[INFO] Config backed up to: /etc/pulse-sensor-proxy/config.yaml.backup.1764788357
[INFO] Pulse server detected at: 192.168.31.31
[INFO] Configuring HTTP mode...
[INFO] Added allowed_source_subnets entry 192.168.31.31/32
[INFO] Added allowed_source_subnets entry 127.0.0.1/32
[INFO] Added allowed_source_subnets entry 192.168.31.11/32
[INFO] HTTP mode configured successfully (using binary)
✓ HTTP mode configured successfully

[INFO] Firewall configuration required:
[INFO]   Allow inbound TCP connections on port 0.0.0.0:8443 from Pulse server
[INFO]   Command: ufw allow from <pulse-server-ip> to any port 0.0.0.0:8443

[INFO] Writing control plane token...
[INFO] Stopping existing service for upgrade...
[INFO] Installing hardened systemd service...
[INFO] Updating service file for HTTP mode...
✓ Service file updated for HTTP mode
[INFO] Enabling and starting service...
[INFO] Waiting for socket...
[INFO] Socket ready at /run/pulse-sensor-proxy/pulse-sensor-proxy.sock
[INFO] Validating HTTP endpoint...
[ERROR] HTTP endpoint validation failed
[ERROR] URL: https://192.168.31.110.0.0.0:8443/health
[ERROR]
[ERROR] Troubleshooting:
[ERROR]   1. Check if port 0.0.0.0:8443 is listening: ss -tlnp | grep 0.0.0.0:8443
[ERROR]   2. Check sensor-proxy logs: journalctl -u pulse-sensor-proxy -n 50
[ERROR]   3. Test manually: curl -k -H 'Authorization: Bearer $TOKEN' https://192.168.31.110.0.0.0:8443/health
[ERROR]
[WARN] Service is running but HTTP endpoint may not be accessible
[WARN] Temperature monitoring may not work properly
[INFO] Installing sensor wrapper script...
✓ Sensor wrapper installed at /opt/pulse/sensor-proxy/bin/pulse-sensor-wrapper.sh
[INFO] Installing cleanup system...
[INFO] Cleanup script installed
[INFO] Cleanup system installed and enabled
[INFO] Configuring proxy SSH access to cluster nodes...
[INFO] Proxy public key: ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIO3cVMvRwLsI+1...

[rcourtman]

Good find on the IPv6 binding issue. Pushed a fix (7535f1f) for the --http-addr URL construction bug.

The installer now correctly extracts the port from both :8443 and 0.0.0.0:8443 formats, so URLs will be properly formed as https://192.168.31.11:8443 instead of the mangled https://192.168.31.110.0.0.0:8443.

For the SSH issues you're seeing in the logs (permission denied for /etc/pve/priv/known_hosts and DNS failures for proxmox2/proxmox3), these are expected when running in standalone HTTP mode on a single node. The proxy is trying to SSH to other cluster nodes but:

1. The pulse-sensor-proxy user doesn't have access to Proxmox's known_hosts file (expected for non-root service account)
2. The hostnames proxmox2, proxmox3 aren't resolvable from that node (they need FQDNs or /etc/hosts entries)

For your single-node setup, these warnings can be ignored - the proxy falls back to direct sensor access (Temperature data fetched successfully via HTTP node=proxmox1).

To get temperatures for all three nodes, you'd need to either:

• Install the sensor-proxy on each node separately (standalone HTTP mode)
• Or configure SSH access between nodes with proper DNS/hosts resolution

The fix for --http-addr will be in the next release, or you can use the installer from main:

curl -fsSL https://raw.githubusercontent.com/rcourtman/Pulse/main/scripts/install-sensor-proxy.sh | \
  sudo bash -s -- --standalone --http-mode --pulse-server http://pulse.kubernetes.home --http-addr 0.0.0.0:8443

[rcourtman]

Also pushed 95ca0c0 - the sensor-proxy now defaults to 0.0.0.0:8443 instead of :8443, which should prevent the IPv6-only binding issue you encountered.

With this fix you shouldn't need to pass --http-addr at all - the default will bind to IPv4.

[zkutasi]

I tried to use the this commit, but I have observed that the Verification part of the script still uses the borked IP:port, so you might have missed a few places with the fix.

Also, even though I specify 0.0.0.0:8443 in the config.yaml, the service still bound only to IPv6. I have tried to check if this is something Proxmox-based, but I have both IPv4 and IPv4 bound ports:

root@proxmox1:~# ss -tulpn -4
Netid      State       Recv-Q      Send-Q           Local Address:Port            Peer Address:Port      Process
udp        UNCONN      0           0                      0.0.0.0:111                  0.0.0.0:*          users:(("rpcbind",pid=678,fd=5),("systemd",pid=1,fd=119))
udp        UNCONN      0           0                    127.0.0.1:323                  0.0.0.0:*          users:(("chronyd",pid=856,fd=5))
udp        UNCONN      0           0                192.168.31.11:5405                 0.0.0.0:*          users:(("corosync",pid=1029,fd=28))
tcp        LISTEN      0           128                  127.0.0.1:9127                 0.0.0.0:*          users:(("pulse-sensor-pr",pid=3455190,fd=7))
tcp        LISTEN      0           4096                 127.0.0.1:12345                0.0.0.0:*          users:(("alloy",pid=795,fd=7))
tcp        LISTEN      0           4096                 127.0.0.1:85                   0.0.0.0:*          users:(("pvedaemon worke",pid=3448646,fd=6),("pvedaemon worke",pid=3444272,fd=6),("pvedaemon worke",pid=3437743,fd=6),("pvedaemon",pid=1079,fd=6))
tcp        LISTEN      0           100                  127.0.0.1:25                   0.0.0.0:*          users:(("master",pid=1019,fd=13))
tcp        LISTEN      0           128                    0.0.0.0:22                   0.0.0.0:*          users:(("sshd",pid=882,fd=3))
tcp        LISTEN      0           4096                   0.0.0.0:111                  0.0.0.0:*          users:(("rpcbind",pid=678,fd=4),("systemd",pid=1,fd=118))
root@proxmox1:~# ss -tulpn -6
Netid      State       Recv-Q      Send-Q           Local Address:Port            Peer Address:Port      Process
udp        UNCONN      0           0                         [::]:111                     [::]:*          users:(("rpcbind",pid=678,fd=7),("systemd",pid=1,fd=121))
udp        UNCONN      0           0                        [::1]:323                     [::]:*          users:(("chronyd",pid=856,fd=6))
tcp        LISTEN      0           100                      [::1]:25                      [::]:*          users:(("master",pid=1019,fd=14))
tcp        LISTEN      0           4096                         *:9633                       *:*          users:(("smartctl_export",pid=809,fd=3))
tcp        LISTEN      0           4096                         *:9100                       *:*          users:(("node_exporter",pid=800,fd=3))
tcp        LISTEN      0           128                          *:8443                       *:*          users:(("pulse-sensor-pr",pid=3455190,fd=8))
tcp        LISTEN      0           4096                         *:3128                       *:*          users:(("spiceproxy work",pid=3340019,fd=6),("spiceproxy",pid=1137,fd=6))
tcp        LISTEN      0           128                       [::]:22                      [::]:*          users:(("sshd",pid=882,fd=4))
tcp        LISTEN      0           4096                      [::]:111                     [::]:*          users:(("rpcbind",pid=678,fd=6),("systemd",pid=1,fd=120))
tcp        LISTEN      0           4096                         *:8006                       *:*          users:(("pveproxy worker",pid=3453690,fd=6),("pveproxy worker",pid=3450935,fd=6),("pveproxy worker",pid=3450842,fd=6),("pveproxy",pid=1112,fd=6))
root@proxmox1:~#

And indeed this also shows port 8443 is only bound on IPv6.
I have this config YAML now:

root@proxmox1:~# cat /etc/pulse-sensor-proxy/config.yaml
# Managed by pulse-sensor-proxy config CLI
allow_idmapped_root: true
allowed_idmap_users:
    - root
allowed_nodes_file: /etc/pulse-sensor-proxy/allowed_nodes.yaml
allowed_peer_uids:
    - 1000
allowed_source_subnets:
    - 192.168.31.31/32
    - 127.0.0.1/32
    - 192.168.31.11/32
    - 192.168.31.0/24
http_auth_token: w6t24ZPWO70l8naU9zh7PB4rnvlgUH2O6Cy1Rv7vV5Y=
http_enabled: true
http_listen_addr: 0.0.0.0:8443
http_tls_cert: /etc/pulse-sensor-proxy/tls/server.crt
http_tls_key: /etc/pulse-sensor-proxy/tls/server.key
metrics_address: 127.0.0.1:9127
pulse_control_plane:
    refresh_interval: 60
    token_file: /etc/pulse-sensor-proxy/.pulse-control-token
    url: http://pulse.kubernetes.home
rate_limit:
    per_peer_burst: 10
    per_peer_interval_ms: 333
root@proxmox1:~#

I consulted LLMs also, according to them, it is like a Go issue, that they default to IPv6 if available. Could you check how binding to IPv4 could be enforced on your side?

[rcourtman]

Good catch on the IPv6 binding issue. Pushed a fix (77e59af).

The root cause was that Go's net.Listen("tcp", "0.0.0.0:8443") doesn't guarantee IPv4-only binding on dual-stack systems with net.ipv6.bindv6only=1. The fix explicitly uses tcp4 network type when the address is IPv4.

The fix will be in the next release. In the meantime, you can test by downloading from main:

# On Proxmox host
sudo systemctl stop pulse-sensor-proxy
curl -L -o /tmp/pulse-sensor-proxy https://github.com/rcourtman/Pulse/releases/latest/download/pulse-sensor-proxy-linux-amd64
# Or build from source if latest release doesn't have it yet

# After next release, re-run installer:
curl -fsSL https://raw.githubusercontent.com/rcourtman/Pulse/main/scripts/install-sensor-proxy.sh | \
  sudo bash -s -- --standalone --http-mode --pulse-server http://pulse.kubernetes.home

After restarting, ss -tlnp -4 | grep 8443 should now show 0.0.0.0:8443 bound on IPv4.

Regarding the install script URL construction bug you mentioned earlier - that was addressed in a previous commit but if you're still seeing issues with the verification part, let me know the specific output.

[zkutasi]

Thanks... trying the latest version with this one-liner usually:

curl -fsSL https://raw.githubusercontent.com/rcourtman/Pulse/refs/heads/main/scripts/install-sensor-proxy.sh | sudo bash -s -- --standalone --http-mode --pulse-server http://pulse.kubernetes.home

But I see you modified the binary now, so I will wait a few iterations. No problem.

About the issue still there in the script, I was not pointing out where, right:

root@proxmox1:~# curl -fsSL https://raw.githubusercontent.com/rcourtman/Pulse/refs/heads/main/scripts/install-sensor-proxy.sh | sudo bash -s -- --standalone --http-mode --pulse-server http://pulse.kubernetes.home
[INFO] Installing pulse-sensor-proxy for standalone/Docker deployment
[INFO] Setting up installation directories...
[INFO] Downloading pulse-sensor-proxy-linux-amd64 from GitHub release v4.36.2...
[INFO] Downloading pulse-v4.36.2-linux-amd64.tar.gz to extract pulse-sensor-proxy...
[INFO] Binary installed to /opt/pulse/sensor-proxy/bin/pulse-sensor-proxy
[INFO] Setting up service directories with proper ownership...
[INFO] Detected pulse-sensor-proxy v4.36.2 (allowed_nodes_file supported)
Migration complete: inline allowed_nodes moved to file
✓ Migration complete: inline allowed_nodes moved to file

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  HTTP Mode Setup (External PVE Host)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

[INFO] Testing connection to Pulse server...
✓ Pulse server is reachable
[INFO] Existing pulse-sensor-proxy detected on port 8443 - will refresh configuration
[INFO] Generating self-signed TLS certificate...
✓ TLS certificate generated (SHA256: D4:31:4F:68:27:6B:E4:C4:DD:6B:2A:9D:92:67:0C:8B:A3:83:14:24:6A:DC:C8:23:33:F4:CF:F2:FB:38:11:8D)
[INFO] Proxy will be accessible at: https://192.168.31.11:8443
[INFO] Registering temperature proxy with Pulse at http://pulse.kubernetes.home...
✓ Registered successfully
[INFO] Config backed up to: /etc/pulse-sensor-proxy/config.yaml.backup.1764879479
[INFO] Pulse server detected at: 192.168.31.31
[INFO] Configuring HTTP mode...
[INFO] Added allowed_source_subnets entry 192.168.31.31/32
[INFO] Added allowed_source_subnets entry 127.0.0.1/32
[INFO] Added allowed_source_subnets entry 192.168.31.11/32
[INFO] HTTP mode configured successfully (using binary)
✓ HTTP mode configured successfully

[INFO] Firewall configuration required:
[INFO]   Allow inbound TCP connections on port 8443 from Pulse server
[INFO]   Command: ufw allow from <pulse-server-ip> to any port 8443

[INFO] Writing control plane token...
[INFO] Stopping existing service for upgrade...
[INFO] Installing hardened systemd service...
[INFO] Updating service file for HTTP mode...
✓ Service file updated for HTTP mode
[INFO] Enabling and starting service...
[INFO] Waiting for socket...
[INFO] Socket ready at /run/pulse-sensor-proxy/pulse-sensor-proxy.sock
[INFO] Validating HTTP endpoint...
[ERROR] HTTP endpoint validation failed
[ERROR] URL: https://192.168.31.110.0.0.0:8443/health
[ERROR]
[ERROR] Troubleshooting:
[ERROR]   1. Check if port 8443 is listening: ss -tlnp | grep 8443
[ERROR]   2. Check sensor-proxy logs: journalctl -u pulse-sensor-proxy -n 50
[ERROR]   3. Test manually: curl -k -H 'Authorization: Bearer $TOKEN' https://192.168.31.110.0.0.0:8443/health
[ERROR]
[WARN] Service is running but HTTP endpoint may not be accessible
[WARN] Temperature monitoring may not work properly
[INFO] Installing sensor wrapper script...
✓ Sensor wrapper installed at /opt/pulse/sensor-proxy/bin/pulse-sensor-wrapper.sh
[INFO] Installing cleanup system...
[INFO] Cleanup script installed
[INFO] Cleanup system installed and enabled
[INFO] Configuring proxy SSH access to cluster nodes...
[INFO] Proxy public key: ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIAYRIFfepyOo3P...

After examining the script, the missing point is at

HTTP_CHECK_URL="https://${PRIMARY_IP}${HTTP_ADDR}/health"

Which missed the HTTP_ADDR parsing-replacement you applied to other parts it seems. It is just the verification, so it did not cause any other issue.

[zkutasi]

By the way:
URL of https://github.com/rcourtman/Pulse/releases/latest/download/pulse-sensor-proxy-linux-amd64 cannot be found. I have stumbled upon this URL myself in the docs, but it no longer works it seems.

Also, I have seen your helm chart stopped at 4.33.1: https://github.com/rcourtman/Pulse/actions/workflows/publish-helm-chart.yml
While the app is already at 4.36.2.