Going to try if this works, thanks!
The logs says:
2023-08-14 16:34:08.902 DEBUG (MainThread) [custom_components.ssh] frigatessh: /sbin/shutdown -r now => , [‘Call to Reboot failed: Interactive authentication required.’], 1
It does the same from shell. I’ll see if you got your command right.
— edit
Right!!
I now understand. This file creates the possibility to execute an elevated command without entering the password. I’ll edit the command in the integration. thanks
How do you get a web terminal?
Hi
I have installed the integration for connecting to a windows device via SSH. The integration connected well but all sensors are in unknown state except Network Status and SSH Status, that show Connected.
Additionally the Power command is disabled. I think this is on purpose as I read the documentation but I do know how to enable it.
Could you help me?
Adrian,
This is fantastic work! Thanks!
Are you able to make it available alongside regular HA integrations, as in via Submit your work | Home Assistant Developer Docs …?
Thanks
Alex
Great package ! Thanks !
I created a sensor command to view available update through apt:
- command: apt-get -s -o Debug::NoLocking=true upgrade |grep -c ^Inst
scan_interval: 1440
sensors:
- type: number
icon: mdi:download-box-outline
name: Available updates
key: available_updates
I also created a action command to update my packages:
- command: sudo /usr/bin/apt-get update; sudo /usr/bin/apt-get -y upgrade
name: Packages update
key: packages_update
icon: mdi:arrow-down-bold-hexagon-outline
I must be missing something REALLY basic… Where do you define commands/sensors? What file and where?
Update: ok, it’s the CONFIGURATION button… duh! Hopefully this will help someone else.
But most of it is working now:
On to the next challenge…
I also have a Docker container running a voice inference server (that uses the GPU).
I can get lots of info out of the log in real time using this command
docker logs --follow 6c3f7e83827b
Now the hard part… I want to extract the variable length sentence
“Tell me a story that takes lots of GPU cycles to infer.” from the logs:
[2023-12-08 23:39:52 +0000] [99] [DEBUG] WHISPER: ASR transcript: Tell me a story that takes lots of GPU cycles to infer.
[2023-12-08 23:39:52 +0000] [99] [DEBUG] WHISPER: Inference took 593.489 ms
[2023-12-08 23:39:52 +0000] [99] [DEBUG] WHISPER: Inference speedup: 9x
The beginning delimiter is “transcript:” and the end is [ on the next line.
I’ve tried piping the output of the ‘docker logs’ command to get the Inference speedup number with
docker logs --tail 1 6c3f7e83827b | awk '{ print substr($0, length($0)-2) }'
But, it doesn’t seem like that command allows piping. I also tried sending it to a file with
docker logs --tail 1 6c3f7e83827b > file.out
but that file was empty.
If I can find the right command to extract the info from the last 3 lines in the log and isolate the info when running the command, I can create sensors that show the values. Heres an example of the command that isolates the GPU Memory used:
nvidia-smi --query-gpu=memory.used --format=csv,noheader | awk '{ print substr( $0, 1, length($0)-4 ) }
So, my linux skills are lacking and I’m at an impasse… for now 
Whew. brutal. “Docker log” doesnt output to stdout, so you have to jump thru hoops…
- command: >-
docker logs --tail 3 6c3f7e83827b 2>&1 |awk 'sub(/.*script\: */,""){f=1}
f{if ( sub(/ *\[.*/,"") ) f=0; print}'
scan_interval: 5
timeout: 2
sensors:
- type: text
maximum: 200
name: STT Result
key: STT_result
icon: mdi:chart-line
- command: >-
docker logs --tail 2 6c3f7e83827b 2>&1 |grep -oP '(?<=took).*(?=ms)'| awk
'{$1=$1};1'
scan_interval: 5
sensors:
- type: number
unit_of_measurement: ms
name: STT Infer time
key: STT_time
icon: mdi:chart-line
- command: >-
docker logs --tail 1 6c3f7e83827b 2>&1 |awk '{ print substr($0,
length($0)-2) }'
scan_interval: 5
timeout: 2
sensors:
- type: text
maximum: 200
name: STT Infer Speed up
key: STT_speedup
icon: mdi:chart-line
and the final result looks like this. I’m very happy w/ this!
Hopefully this helps someone else.
I honestly don’t understand how come this integration doens’t get tons and tons of stars in HACS. It is by far the best way to push SSH commands imo!
Hi,
This is my 2 cents on nvidia gpu and some other sensors. I use docker with nvidia toolkit.
- command: nvidia-smi --query-gpu=temperature.gpu --format=csv,noheader
scan_interval: 60
sensors:
- type: number
name: GPU Temperature
key: temperature
separator: ","
unit_of_measurement: °C
- command: nvidia-smi --query-gpu=utilization.gpu --format=csv,noheader,nounits
scan_interval: 60
sensors:
- type: number
name: GPU load
key: gpu_load
unit_of_measurement: "%"
- command: nvidia-smi --query-gpu=memory.used --format=csv,noheader,nounits
scan_interval: 60
sensors:
- type: number
name: GPU ram used
key: used_memory
unit_of_measurement: MiB
- command: nvidia-smi --query-gpu=memory.total --format=csv,noheader,nounits
scan_interval: 60
sensors:
- type: number
name: GPU ram total
key: total_memory
unit_of_measurement: MiB
- command: nvidia-smi --query-gpu=memory.free --format=csv,noheader,nounits
scan_interval: 60
sensors:
- type: number
name: GPU ram free
key: free_memory
unit_of_measurement: MiB
- command: nvidia-smi --query-gpu=fan.speed --format=csv,noheader,nounits
scan_interval: 60
sensors:
- type: number
name: GPU fan
key: fan_speed
unit_of_measurement: "%"
- command: nvidia-smi --query-gpu=utilization.gpu --format=csv,noheader,nounits
sensors:
- type: number
name: GPU utilization
key: utilization_gpu
unit_of_measurement: "%"
- command: nvidia-smi --query-gpu=utilization.memory --format=csv,noheader,nounits
sensors:
- type: number
name: GPU mem utilization
key: memory_utilization
unit_of_measurement: "%"
- command: who --count | awk -F "=" 'NR>1 {print $2}'
scan_interval: 60
sensors:
- type: number
name: Logged in users
icon: mdi:account-multiple
- command: cat /proc/uptime | awk '{print $1}'
scan_interval: 180
sensors:
- type: number
name: Uptime
unit_of_measurement: d
value_template: "{{ value | float // 86400 }}"
icon: mdi:server
- command: who | awk '{print $1}'
scan_interval: 60
sensors:
- type: text
name: Logged user
icon: mdi:account
I have added your sensors. Thanks for that! Haven’t been able to test the update button yet since there are no updates ;-). Unfortunately I cannot seem to get the sensor going. It keeps reporting Unknown. When I execute the command inside the regular SSH putty window it reports 0. Any clue? Using Bookworm debian on raspi 4.
This is honestly exceptional work. I couldn’t figure it out for the life of me how to send commands to turn off an external display (remote lovelace dashbaord connect to a pi). Your integration helped solve that in literal minutes. Now I can attach a motion sensor to turn on/off a monitor. Amazing stuff! Thanks so much
Can you share your code here. Maybe someone will find that useful.
To answer my own question, it reports unknown when there are No updates available. When there are, it’ll report a number. It would be great if it would report 0 instead of ‘unknown’, but don’t know how to do that.
Hi,
it would nice to also have the “uptime” available as a sensor.
BR
Ben
If you add the following to the configuration, it’ll show up as a sensor:
Hey hopefully this integration is still maintained, I just moved over to this from SSH Sensor, which hasn’t been maintained for three years and stopped working with 2024.7.0.
Anyway I had to slightly rewrite one of my sensors (uptime) moving from that unmaintained sensor. Here it is, if it helps!
- command: uptime -s
scan_interval: 60
sensors:
- type: text
name: uptime
key: uptime
value_template: >-
{%- set line = value.split("\r\n")[0] -%}
{%- set lastboot = as_timestamp(strptime(line, "%Y-%m-%d %H:%M:%S")) -%}
{%- set now = as_timestamp(now()) -%}
{%- set uptime = now - lastboot | int -%}
{%- set years = uptime // 31536000 -%}
{%- set months = (uptime % 31536000) // 2592000 -%}
{%- set days = (uptime % 2592000) // 86400 -%}
{%- set hours = (uptime % 86400) // 3600 -%}
{%- set minutes = (uptime % 3600) // 60 -%}
{{ '%dy ' % years if years else '' }}{{ '%dm ' % months if months else '' }}{{ '%dd ' % days if days else '' }}{{ '%dh ' % hours if hours else '' }}{{ '%dm ' % minutes if minutes else '' }}
hello everyone. tell me how to integrate the nload command and get a graph of the network connection speed
Hello, I’m trying to setup this integration but I’m facing an issue. when I try with this configuration :
I get Host is offline
I have enabled debug logs but it doesn’t give me much more information:
2024-08-05 19:32:37.296 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: network_interface => None
2024-08-05 19:32:37.297 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: mac_address => None
2024-08-05 19:32:37.297 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: wake_on_lan => None
2024-08-05 19:32:37.297 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: hostname => None
2024-08-05 19:32:37.297 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: machine_type => None
2024-08-05 19:32:37.298 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: os_name => None
2024-08-05 19:32:37.298 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: os_version => None
2024-08-05 19:32:37.298 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: device_name => None
2024-08-05 19:32:37.298 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: device_model => None
2024-08-05 19:32:37.299 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: manufacturer => None
2024-08-05 19:32:37.299 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: serial_number => None
2024-08-05 19:32:37.299 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: cpu_name => None
2024-08-05 19:32:37.299 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: cpu_cores => None
2024-08-05 19:32:37.299 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: cpu_hardware => None
2024-08-05 19:32:37.300 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: cpu_model => None
2024-08-05 19:32:37.300 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: total_memory => None
2024-08-05 19:32:37.300 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: free_memory => None
2024-08-05 19:32:37.300 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: cpu_load => None
2024-08-05 19:32:37.301 DEBUG (SyncWorker_54) [custom_components.ssh.config_flow] 192.168.1.202: processes => None
2024-08-05 19:32:37.304 WARNING (MainThread) [custom_components.ssh.config_flow] Host is offline
I know that the host (192.168.1.202) is not down and that it is reachable from homeassistant because I can ssh to it manually from a terminal when I’m using Home Assistant Community Add-on: SSH & Web Terminal.
Any ideas what the issue could be ?
Hi,
Just wanted to say that this is a great add-on. I have been trying to perform actions and update sensors from my ZFS Server to no avail.
Then I stumbled upon this and had my sensors up and running within minutes.
Fantastic,
Much appreciated