------------------------------------------------------------------
_________ ______________________________
/ _____/__.__. ______ \_____ \__ ___/\__ ___/
\_____ < | |/ ___/ _____/ / \ \| | | |
/ \___ |\___ \ |Sys-QTT|/ \_/. \ | | |
/_______ / ____/____ >"`` \_____\ \_/____| |____|
\/\/ \/ \__> o
System Metrics MQTT Client
------------------------------------------------------------------
Sys-QTT is a light-weight, Python-based system metrics client for monitoring networked devices. It periodically gathers a customisable selection of host machine metrics, and publishes the data to an MQTT broker/server of your choice.
Sys-QTT is based on Sennevds' System Sensors project, and has been developed primarily to use with Home Assistant, based on their MQTT discovery documentation. It should work with other any other MQTT broker/configuration, however, the sensor configuration messages are specific to Home Assistant.
The config/settings for this fork is not compatible with System Sensors. Please follow the instructions below to reeneter the settings from the supplied example directory.
The config.yaml file provides a selection from the following metrics:
- CPU: model, temperature, number of threads and cores, usage %, and current & max clock-speed
- Average Load: 1min, 5min and 15min
- Storage: file-system and mounted volume drive usages
- Memory: physical memory usage, swap usage
- Network: tx/rx rate, local IP, WiFi signal strength and SSID
- OS: hostname, distro name, distro version and pending OS updates
- Hardware: system architecture, board make and model
- Timestamps: last boot, last message received
- Add board make and model sensors
- Comprehensive logging for debug
- Clean up the project file structure
- Add more CPU details
- Complete refactor of the sensors code-base
- Move sensor definitions to
sensor_details.jsonfile - Reduce function calls by defining sensors as static or dynamic
- Move sensors to a run-time dictionary of sensor objects
- Re-organise scripts so clear their readibility
- Move sensor definitions to
- Add ability to create custom sensor (instead of all being hard-coded):
- Build a set of bash script modules that can be called by custom sensors
- Add fields in
sensor_details.jsonfor custom bash calls and string searches
- Tidy Sys-QTT Home Assistant entities:
- Move semi-permanent sensor values (make, model, max, etc.) to attributes(?)
- Utilise Home Assistant v2021.11's Entity Categories
- This will clean up the huge amount of sensors, especially with multiple Sys-QTT deployments
- Install/update script
- Web interface (yeah, I'm not a minimalist)
- Build custom HA integration for managing remote Sys-QTT nodes
Platform: This has only be tested on, and will likely only run on, Linux systems.
- Python 3.8+
- Several Python modules (installed via the supplied
requirements.txtfile)
This project assumes you have an MQTT broker already running. For example:
- (Home Assistant) Mosquitto broker integration
- (Stand-alone) Eclipse-Mosquitto
The following steps will configure the host and perform a test-run of Sys-QTT:
- Clone the repo:
git clone https://github.com/MaxVRAM/Sys-QTT.git && cd Sys-QTT- Ensure the
python3-aptpackage is installed:
sudo apt-get install python3-apt- Install the required modules:
python3 -m pip install -r requirements.txt- Make a copy of the example config file and update it with your MQTT broker details, client ID and client device name:
cp examples/config.yaml config.yaml && nano config.yaml- Test the script within your CLI session, any issues will be clearly logged:
python3 sys-qtt.py --config config.yamlClick here
❯ python3 sys-qtt.py
-----------------------
Sys-QTT starting up...
-----------------------
[•] Importing config.yaml...
[✓] Config file found: /home/maxvram/Sys-QTT/config.yaml
[•] Processing config...
[✓] Config initialised.
[•] Importing sensor properties...
[✓] Sensor properties loaded.
[✓] Config loaded successfully.
[•] Importing sensor configurations...
[✓] Imported 25 sensor properties.
[•] Initialising static sensors...
[✓] Static sensors built.
[•] Checking output of each sensor...
[✓] board_make returned: Intel Corporation
[✓] board_model returned: NUC8BEB
[✓] cpu_arch returned: x86_64
[✓] cpu_model returned: Intel(R) Core(TM) i5-8259U CPU @ 2.30GHz
[✓] cpu_threads returned: 8
[✓] cpu_cores returned: 4
[✓] cpu_max returned: 3.8 GHz
[✓] cpu_clock returned: 2.3 GHz
[✓] cpu_temp returned: 38.0 °C
[✓] cpu_usage returned: 11.9 %
[✓] cpu_load_1m returned: 1.73
[✓] cpu_load_5m returned: 1.09
[✓] cpu_load_15m returned: 1.02
[✓] memory_ram returned: 41.6 %
[✓] memory_swap returned: 97.5 %
[✓] os_hostname returned: NUC
[✓] os_distro returned: Ubuntu 20.04.3 LTS
[✓] os_updates returned: 0
[✓] net_ip returned: 192.168.20.5
[✓] net_tx returned: 0 Kbps
[✓] net_rx returned: 0 Kbps
[✓] last_boot returned: 2021-10-27T01:18:41+11:00
[✓] last_message returned: 2021-11-09T00:24:34.510696+11:00
[✓] disk_system returned: 28.3 %
[✓] disk_storage returned: 0.3 %
[✓] 25 sensors have been commited to the session.
[✓] Local configuration complete.
[•] Attempting to reach MQTT broker at 192.168.20.5 on port 1883...
[✓] MQTT broker responded.
[•] Publishing sensor configurations...
[✓] 25 sensor configs and online status to broker.
[•] Establishing MQTT connection loop...
[✓] Success!
[i] Updated nuc_i5 client on broker with online status.
[•] Adding sensor update job on 30 second schedule...
[✓] [Every 30 seconds do publish_sensor_values() (last run: [never], next run: 2021-11-09 00:25:05)]
------------------------------
Sys-QTT now running on: NUC i5
------------------------------
[•] Sending update sensor payload...
[✓] 25 sensor updates sent to MQTT broker.
[•] 30 seconds until next update...
If the installation and test went well, you can now add a run service for Sys-QTT to run in the background on boot:
- Open and edit the example
sys-qtt.servicefile, changing the user and exec file paths to reflect your configuration:
nano examples/sys-qtt.service- Copy the updated service file to your system service directory:
sudo cp examples/sys-qtt.service /etc/systemd/system- Enable and start the service:
sudo systemctl enable sys-qtt
sudo systemctl start sys-qtt- Check that the service started correctly:
sudo systemctl status sys-qttThe last few lines of the Sys-QTT log output can be viewed via the systemd status command:
sudo systemctl status sys-qttIf this doesn't provide any insights into the issue, stop the systemd service, and run Sys-QTT manually in the CLI:
sudo systemctl stop sys-qtt
python3 ~/Sys-QTT/sys-qtt.pySys-QTT will then start in your CLI and display the log in full detail as it starts up.
Another option is to use MQTT Explorer, which is an excellent tool for monitoring the activity of your MQTT data. If the metrics aren't appearing in your Home Assistant listings, I recommend downloading and connecting MQTT Explorer to your broker.
I have used following custom plugins for lovelace:
- vertical-stack-in-card
- mini-graph-card
- bar-card
Config:
- type: 'custom:vertical-stack-in-card'
title: Deconz System Monitor
cards:
- type: horizontal-stack
cards:
- type: custom:mini-graph-card
entities:
- sensor.deconz_cpu_usage
name: CPU
line_color: '#2980b9'
line_width: 2
hours_to_show: 24
- type: custom:mini-graph-card
entities:
- sensor.deconz_temperature
name: Temp
line_color: '#2980b9'
line_width: 2
hours_to_show: 24
- type: custom:bar-card
entity: sensor.deconz_disk_use
title: HDD
title_position: inside
align: split
show_icon: true
color: '#00ba6a'
- type: custom:bar-card
entity: sensor.deconz_memory_use
title: RAM
title_position: inside
align: split
show_icon: true
- type: entities
entities:
- sensor.deconz_last_boot
- sensor.deconz_under_voltageExample:
