Simply it subscribes to MQTT topics and exposes data published there in way that is easy to read with Prometheus
web.listen-address: [Default: ":2112"] Address on which to expose metrics and web interface.
web.telemetry-path: [Default: "/metrics"] Path under which to expose metrics.
mqtt.host: [Default: "127.0.0.1:1883"] Mqtt host address and port.
mqtt.clientId: [Default: "mqtt_exporter"] Mqtt clientId.
mqtt.username: [Default: ""] Mqtt username.
mqtt.password: [Default: ""] Mqtt password
naming.config: [Default: "/etc/mqtt_exporter/naming.yaml"] File containg naming convertions
metrics.prefix: [Default: "mqtt_exporter"] Prefix for metrics names
log.level: [Default: 2] Log level
cleaner.gauge.timeout: [Default: 0s] Timeout for gauge value cleaner (0 = disabled)
Group1: # "group" attribute of metric
- device: device_name # device name from mqtt message
# used for matching
# "device" attribute of metrics
name: friendly_name # "friendly_name" attribute value
sensors: # list of "sensor_name"
sensor_name: sensor_alias # for this device for every readout from "sensor_name" there will be "sensor_alias" label added
| value | meaning |
|---|---|
| 1 | DEBUG |
| 2 | INFO |
| 3 | WARN |
| 4 | ERROR |
| 5 | OFF |
Prevents dangling metrics in prometheus by removing metrics that were not updated for set time. Metrics will be removed within 5 seconds after cleaner.gauge.timeout of not being updated.
Useful when one of your devices disappears.
First build project:
cd $GOPATH/src/github.com/klaper_/mqtt_data_exporter
go build -o bin/mqtt_data_exporter
Build docker image:
docker build -t mqtt_data_exporter .
Run container:
docker run -p 2112:2112 mqtt_data_exporter --mqtt.host "mqtthost.local:1883"