At the moment, the autodiscovered sensors are not force-updated. That leads to skewed statistics because Homeassistant doesn't update an MQTT sensor when the value is the same and then by extension doesn't update the recorder database nor forwards values e.g. to Influxdb.
That leads to skewed statistics.
Example:
My weight has been constant at 95kg for a week. Then I go out, drink some alcohol. The next morning, my weight is at 98kg. Because Homeassistant only records changed data unless forced otherwise, in the graph it will look like my weight has risen constantly from 95kg to 98kg the whole week, instead of the spike related to the night out.
Solution:
Add force_update: true to the sensor configuration in the autodiscovery. That will make Homeassistant update the value every time it receives one, no matter if it's identical to the previous one.
At the moment, the autodiscovered sensors are not force-updated. That leads to skewed statistics because Homeassistant doesn't update an MQTT sensor when the value is the same and then by extension doesn't update the recorder database nor forwards values e.g. to Influxdb.
That leads to skewed statistics. Example: My weight has been constant at 95kg for a week. Then I go out, drink some alcohol. The next morning, my weight is at 98kg. Because Homeassistant only records changed data unless forced otherwise, in the graph it will look like my weight has risen constantly from 95kg to 98kg the whole week, instead of the spike related to the night out.
Solution: Add
force_update: trueto the sensor configuration in the autodiscovery. That will make Homeassistant update the value every time it receives one, no matter if it's identical to the previous one.