I've been fulfilling my hunger for outgoing MQTT from Hubitat by posting events using the Maker API to Nginx using some string and sticky-tape . It's been solid as a rock, but this new app is far more flexible, and incorporates Home Assistant topics too!
Very impressed, thank you Hubitat crew!
As a bit of project self-promotion, if you didn't already know you can also get Zigbee2MQTT devices in to Hubitat using my Zigbe2MQTT Routing Driver, so everything suddenly ties together very nicely indeed!
I was replacing a couple of MQTT feeds with the new ones from Hubitat, but now get "The entity no longer has a state class" in the Home Assistant Repairs screen.
The MQTT feed needs the correct SensorStateClass declared to store long-term statistics. For most sensors this should just be the addition of "state_class": "measurement" to the MQTT payload. There are also SensorStateClass.TOTAL and SensorStateClass.TOTAL_INCREASING options too.
Just to add to the high praise for the MQTT Exporter it seems to be working very well for me. I have actually managed to use the MQTT Exporter in conjunction with Influxdb v3 and it's MQTT Subscriber Plugin. I am still running "InfluxDB Logger" for a older instance of InfluxDB v2 which is i consider my stable data, butt may not need to going forward with this new setup with the version of influxDB.
Kudo's guys on getting this going that is so light weight.
What does that actually mean. What problems did you experience. There is atleast one issue that gets addresed in the next release. When i first tested MQTT i had a issue that caused my hub to become unresponsive and created a resource issue. Maybe the same problem. It seems to have been triggered by someting happening during change to the app configuration.
Mentioned it above, but Hub went into constant "Elevated Load" state, web interface became unresponsive. Had to use diagnostics page to reboot. Load went from between 0.6-0.9, to constant 1.8-2.0, Zigbee network going offline, devices unresponsive.
Back on v2.4.3.173 and it's solid as a rock again.
That is the effect of you disabling the sensor. Is not really an error. Is more of a warning that state value field was empty and can't be saved in the long term statistics. You should delete the records from the statistics and start over collecting the data when the sensor is active again. If you don't need statistics on the sensor, you can exclude the sensor in your configuration file and you will no longer see that error if you disable the sensor again.