Skip to content

cottephi/raspberry_sensors

BranchesTags

Repository files navigation

Raspberry Sensors

This little project aims at operating various sensors on a Raspberry Pi, store their data in an InfluxDB database, and use Grafana to display them.

Available sensors

Plugging the sensors

Raspberry I2C ports

Raspberry I2C ports

Activate I2C connections on your Raspberry Pi

BME280

Plug the wires on the Raspberry's I2C pins according to this table:

Function Pin (BME280) I2C Pin (Raspberry)
ADDR/MISO GND
SCL/SCK SCL
SDA/MOSI SDA
GND GND
VCC 3.3V/5V
CS not used

(Found here)

Installation

Assuming you plugged everything correctly

Clone the reposotiry, cd in it and compile the code with

go build -o raspberry_sensors.exe ./cmd/raspberry_sensors/

Main program configurations

Copy the following YAML configuration in a config.yml file at the root of the repository:

api:
  # Overriden by env var SERVER_HOST, default is localhost
  host: "localhost"
  # Overriden by env var SERVER_PORT, default is 8080
  port: 8080

database:
  # Overriden by env var DB_HOST, default is http://localhost
  host: http://localhost
  # Overriden by env var DB_PORT, default is 8086
  port: 8086
  # Overriden by env var DB_TOKEN, no default, mandatory
  token: <your influx DB token>

logger:
  # zerolog levels are 'trace', 'debug', 'info', 'warn', 'error', 'fatal' and 'panic'
  level: "info"

To test the sensors code, you do not need to give InfluxDB's token at first. Instead, run

./raspberry_sensors.exe -d -s

The -d option indicates a dry run, which will only output the result in the console, and note write anything to the database. The -s option indicates that the data acquisition should start right away. If everything is normal, you should see:

No logfile path provided, only logging in the console
2024-09-08T11:59:01+02:00 INF cmd/raspberry_sensors/main.go:15 > Configuration:
 - Server Host: http://localhost
 - Server Port: 8080
 - Logger Level: INFO
 - Log file path: 
 - Server URL: http://localhost:8080
No database token given, not writing data to database
2024-09-08T11:59:01+02:00 WRN internal/utils/utils.go:25 > No InfluxDB token provided. Not starting Database.
2024-09-08T11:59:01+02:00 INF internal/utils/utils.go:42 > Creating Sensors...
2024-09-08T11:59:01+02:00 INF internal/utils/utils.go:46 > ...ok
2024-09-08T11:59:01+02:00 INF internal/utils/utils.go:52 > Starting sensor BME280...
2024-09-08T11:59:01+02:00 INF internal/utils/utils.go:59 > ...ok
2024-09-08T11:59:01+02:00 INF internal/utils/utils.go:66 > Starting server...
2024-09-08T11:59:01+02:00 INF internal/utils/utils.go:69 > ...ok
2024-09-08T11:59:01+02:00 INF internal/api/server.go:61 > Listening/Serving on localhost:8080
2024-09-08T11:59:01+02:00 INF internal/utils/utils.go:77 > Sending data acquisition start signal right away
2024-09-08T11:59:01+02:00 INF internal/sensors/sensor.go:84 > Starting data acquisition of sensor BME280
2024-09-08T11:59:01+02:00 INF internal/sensors/bme280.go:17 > Temperature: 25.87°C, Pressure: 995.90 hPa, Humidity: 52.23%
2024-09-08T11:59:02+02:00 INF internal/sensors/bme280.go:17 > Temperature: 25.88°C, Pressure: 995.88 hPa, Humidity: 52.18%
2024-09-08T11:59:03+02:00 INF internal/sensors/bme280.go:17 > Temperature: 25.88°C, Pressure: 995.91 hPa, Humidity: 52.16%

If all is fine, execute ./setup_log_dirs.sh and configure the database.

Database configuration

Go to http://localhost:8086 and login to InfluxDB. Then do the following

  • Create an organisation named "raspberry"
  • Create a bucket named "seconds" with a retention policy of 2 days
  • Create a bucket named "daily". Choose the retention policy (I did not use one)
  • Import the task found in internal/database/hourly_aggregation.json in InfluxDB Tasks.
  • Create a token for your sensor, or use the admin one (not recommended). It only needs read/write access to the 'seconds' bucket.
  • Create a token for Grafana, with read access to both buckets

Modify the /etc/systemd/system/influxd.service file by adding the line

Environment="INFLUXD_HTTP_BIND_ADDRESS=0.0.0.0:8086"

in the Service section. Reload systemctl and restart the service:

sudo systemctl daemon-reload
sudo systemctl restart influxdb

You can now configure the services.

Service configuration

To make the program start with your Raspberry Pi, create a Systemd service :

sudo nano /etc/systemd/system/raspberry_sensors.service

with the following content

[Unit]
Description=Raspberry Sensors Go Program
After=influxdb.service

[Service]
WorkingDirectory=<your home>/projects/raspberry_sensors
ExecStart=<your home>/projects/raspberry_sensors/raspberry_sensors.exe -s
Restart=always
User=<your user name>

[Install]
WantedBy=multi-user.target

and a second one

sudo nano /etc/systemd/system/docker-compose-monitoring.service

with the following content

[Unit]
Description=Docker Compose Stack for Grafana, Promtail, Loki
Requires=docker.service
After=docker.service

[Service]
WorkingDirectory=<your home>/projects/raspberry_sensors
ExecStart=/usr/bin/docker compose up
ExecStop=/usr/bin/docker compose down
Restart=always
User=<your user name>

[Install]
WantedBy=multi-user.target

Run then

sudo systemctl daemon-reload
sudo systemctl enable docker-compose-monitoring.service
sudo systemctl enable raspberry_sensors.service
sudo systemctl start docker-compose-monitoring.service
sudo systemctl start raspberry_sensors.service

Grafana configuration

The program can also write logs to a file that Grafana/Loki will process. You already ran setup_log_dirs.sh, that creates the loki directory in your home folder, and created a local .env file that the program will detect, and from which the path where the logs should be written are read. It is $HOME/logs/raspberry_sensors. the directory is created when you first run the program.

If you correctly set up your two services, you should be able to check the logs of Grafana, Loki and Promtail with docker logs grafana/loki/promtail.

If all is good, log into Grafana (http://localhost:3000) with admin:admin. Then go to Dashboard -> New -> Import and use the file internal/grafana/Logs-dashboard.json. You might have to setup you data source to Loki first.

The log files are rotating files of at most 1GB each, keeping at most the last 5 log files for at most 7 days.

To graph the sensors results, in Grafana, set up the data source for InfluxDB (see this link)

You can import the dashboards internal/grafana/BME280-daily-aggregation-dashboard.json and internal/grafana/BME280-Data-dashboard.json.

BONUS: Show the raspberry Disk, CPU and memory usage in InfluxDB

To monitor you Raspberry, you can configure InfluxDB to use Telegraf as a data source. First, install telegraf, then go to your InfluxDB dashboard (http://localhost:8086) create a bucket named "metrics" (I chose a 7 days retention policy, but that is up to you). Then go to Sources -> telegrag -> create configuration. Select the bucket "metrics", "telegraf internal" as source and paste the following:

# Read metrics about CPU usage
[[inputs.cpu]]
  percpu = true
  totalcpu = true
  collect_cpu_time = false
  report_active = true

# Read metrics about memory usage
[[inputs.mem]]

# Read metrics about disk usage
[[inputs.disk]]
  ignore_fs = ["tmpfs", "devtmpfs", "overlay"]

Note the token that you are provided and the command to start telegraf.

Then, edit the telegraf Systemd service:

sudo nano /lib/systemd/system/telegraf.service

Add the Environment=<your token> line in the [Service] section and replace the --config option with the one provided by InfluxDB UI. Change the After and Wants values to be influxdb.service. My final configuration looks like that for example:

[Unit]
Description=Telegraf
Documentation=https://github.com/influxdata/telegraf
After=influxdb.service
Wants=influxdb.service

[Service]
Type=notify
NotifyAccess=all
EnvironmentFile=-/etc/default/telegraf
User=telegraf
ImportCredential=telegraf.*
ExecStart=/usr/bin/telegraf --config http://192.168.1.15:8086/api/v2/telegrafs/0da2cd81a15fe000
Environment="INFLUX_TOKEN=***"
ExecReload=/bin/kill -HUP $MAINPID
Restart=on-failure
RestartForceExitStatus=SIGPIPE
KillMode=mixed
LimitMEMLOCK=8M:8M
PrivateMounts=true

[Install]
WantedBy=multi-user.target

Restart your telegraf service:

sudo systemctl disable telegraf
sudo systemctl daemon-reload
sudo systemctl enable telegraf
sudo systemctl restart telegraf.service

In the InfluxDB UI, check that you have data in your bucket metrics. If so, you can create a new Dashboard in InfluxDB UI by uploading the one in internal/database/metrics.json.

Useful commands

Empty a bucket in the database

influx delete --bucket <bucket> --predicate '_measurement="environment_bme280"' --start '1970-01-01T00:00:00Z' --stop $(date +"%Y-%m-%dT%H:%M:%SZ") --org raspberry --token <admin token>

Stop data acquisition without killing the program

wget -qO- localhost:8080/sensors/stop

Start data acquisition (if you stopped it with the previous command or did not start the program with the -s flag)

wget -qO- localhost:8080/sensors/start

Kill the program

wget -qO- localhost:8080/sensors/kill

Note that the API routes can be available from any computer on your local network if you know your Raspberry Pi's IP address.

About

No description, website, or topics provided.

Resources

Readme
Activity

Stars

0 stars

Watchers

1 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

No packages published

Languages