Over the weekend, I setup wfudplistener to print the Tempest hub’s UDP obs_st data to a file. The log data is then shipped in realtime to a LogScale (log aggregator) instance via a log shipper where it can be parsed and displayed. The result is here.
Id suggest you consider using a websocket vs UDP. UDP is easy to miss packets and the websocket has more data as it includes the data that is adjusted by WeatherFlow like lightning strike distance and NearCast.
Hi there. Thank you all for the feedback. The listener is wired and currently in the past 7 days I’ve dropped one packet. Since the interval is 1 minute, and it’s weather, I’m not as concerned as I would be with traditional log data that is mission critical. Cheers!
The websocket still includes the non hacked data, it just adds the hacked data on top of it.
One thing to note here, also, the lightning strike data from your sensor is not the distance from the lightning strike. It is the caluclated distance to the front of the storm. It can detect a lightning strike 25 miles away, but if there was a recent strike 1 mile away, the UDP data will report 1, not 25. If lightning data is important, you’ll want the “hacked” data anyways as it reports actual strike distance.
How do you like LogScale? I have been trying out GrayLog and it’s heavy on the MongoDB and Open or Elastic Search.
I also capture UDP packets because why should I have to make calls to a vendors API vs. the data being local and use WeeWx and the UDP driver for the method to collect the packets. I also directly republish the UDP data on a MQTT topic so I can subscribe to data from other devices like home automation.