My power monitoring setup/scripts using TP-Link HS110, Python, InfluxDB, Grafana
This was installed on a raspberry pi. Some commands may vary depending on raspbian/debian/ubuntu.
These are using TP-Link HS110 devices.
Warning: The HS100 looks similar but it doesn't have power monitoring. Check before you buy.
I just set them up in the normal way using the Kasa app, although I think you can bootstrap them in isolated mode.
From https://docs.influxdata.com/influxdb/v1.8/introduction/install/
wget -qO- https://repos.influxdata.com/influxdb.key | sudo apt-key add -
source /etc/os-release
echo "deb https://repos.influxdata.com/debian $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/influxdb.list
sudo apt-get install influxdb
In my case influx service was started automatically, if not follow steps from above link for service configuration.
Then based on https://docs.influxdata.com/influxdb/v1.8/introduction/get-started/ login to influxdb and create the database.
In my case and used in scripts here it's database EnergyDB.
me@host:~ $ influx -precision rfc3339
Connected to http://localhost:8086 version 1.8.2
InfluxDB shell version: 1.8.2
> CREATE DATABASE EnergyDB
> SHOW DATABASES
name: databases
name
----
_internal
EnergyDB
> exit
From https://grafana.com/docs/grafana/latest/installation/debian/ follow "To install the latest OSS release":
sudo apt-get install -y apt-transport-https
sudo apt-get install -y software-properties-common wget
wget -q -O - https://packages.grafana.com/gpg.key | sudo apt-key add -
echo "deb https://packages.grafana.com/oss/deb stable main" | sudo tee -a /etc/apt/sources.list.d/grafana.list
sudo apt-get update
sudo apt-get install grafanaLogin to Grafana at http://HOSTNAME:3000/ where HOSTNAME = IP where you installed it
Configuration -> Data Sources -> Add data source -> InfluxDB
Set URL. I set it to http://HOSTNAME:8086 since it's running on same server as Grafana. Add
To communicate with and get data from the HS110 plugs we use https://github.com/GadgetReactor/pyHS100 To send data to InfluxDB we use https://github.com/influxdata/influxdb-python
pip install pyHS100
pip install influxdbDiscover devices using either pyhs100 command:
~ $ pyhs100
No host name given, trying discovery..
Discovering devices for 3 seconds
== DeviceLabel - HS110(EU) ==
Device state: ON
Host/IP: 192.168.XXX.XXX
On since: 2020-09-14 ........
LED state: True
== Generic information ==
Time: 2020-09-25 ....
Hardware: 4.0
Software: 1.0.4 Build 191111 Rel.143500
MAC (rssi): AB:CD:EF:AC:CD:EF (-43)
Location: {'latitude': XXXX, 'longitude': XXXX}
== Emeter ==
== Current State ==
{'power_mw': 4155, 'voltage_mv': 237051, 'current_ma': 214, 'total_wh': 412}
or using python snippet:
from pyHS100 import Discover
for dev in Discover.discover().values():
print(dev)Note the Device IP addresses like 192.168.XXX.XXX
- Copy hs110toinflux.py to somewhere on the system.
- for example in user's home folder so it's
/home/USERNAME/hs110_influx_feed.py
- for example in user's home folder so it's
- set IP/Database of influx client
client = InfluxDBClient(host='192.168.XXX.XXX', database="EnergyDB") - set HS110 device IPs in
plug_ips = ["192.168.XXX.XXX", "192.168.XXX.XXX", "192.168.XXX.XXX"]
create /home/USERNAME/.config/systemd/user/hs110.service containing:
[Unit]
AssertPathExists=/home/USERNAME/
Description=hs110 data collection
[Service]
WorkingDirectory=/home/USERNAME
ExecStart=/usr/bin/python3 /home/USERNAME/hs110_influx_feed.py
Restart=always
[Install]
WantedBy=default.target
Run commands:
# enable linger for user so service can start at boot time without interactive login
sudo loginctl enable-linger USERNAME
# reload service definitions
systemctl --user daemon-reload
# start the service
systemctl --user start hs110
# enable automatic startup of service on boot
systemctl --user enable hs110
Data should be fed into InfluxDB now.
- Create->Dashboard
- Add new Panel
- in Query dropdown select EnergyDB data source
- in FROM change "select measurement" to "power"
- in SELECT change field(value) to field(power)
- you should already start seeing graph data showing milliwatts
- add more graphs, group by device, ask someone who knows grafana better
New 2020-09-30
I also want to gather long term data by day/week/month and compare it to the power bill. And this presents some problems:
- my InfluxDB feeder may fail (from experience, saw it happen) sometimes or miss some data points
- Averaging even over measurements made every second could give odd results
- too many data points to work with, totals per day would be good enough
Thankfully the HS110 plug does store total power usage per day on the plug and this is available via pyHS100. As far as I know it stores last 30 days with totals per day and also aggregate monthly numbers.
We want to be able to look up a certain day's measurements even from a few months ago, so we're just going to periodically save the data in our own small SQLite database.
- pre-requisites:
pip install jinja2if not existing - copy hs110_daily_gather.py somewhere on the system
- also copy daily_template.html in the same folder as the python script
- set HS110 device IPs in
plug_ips = ["192.168.XXX.XXX", "192.168.XXX.XXX", "192.168.XXX.XXX"] - make the script executable if needed using
chmod +x hs110_daily_gather.py - run
./hs110_daily_gather.pywhich will do the following- it will create a
daily_power_usage.dbfile if not existing. This is the SQLite database - it will gather daily usage data from the HS110 plugs for the current month and previous month and update them in the database
- it will render the
daily_template.htmlasdaily_rendered.htmlfile. This is a basic bootstrap styled table as can be seen below with all the power usage totals
- it will create a
- set a cronjob to run the
hs110_daily_gather.pyscript at least daily or more frequently if you wish the current day's total usage to be more up to date - Optional: I also created a symlink to serve it via my webserver and plugged it into Grafana using the AJAX Panel Plugin ( https://grafana.com/grafana/plugins/ryantxu-ajax-panel )
