- Posted on
Sun Jul 30, 2017 8:07 pm
-
smurfless1
offline
-
- Posts: 38
- Joined: Sep 23, 2015
- Location: Round Rock, TX
For those not familiar, Kafka is mostly like a DVR for log entries. It has channels and a limit of how many to keep. Then scripts/software can subscribe to a channel (called a topic) for either realtime or catch-up processing. So, you can store "1 TB of my Indigo info" or "1 year of Indigo info". And it lets clients keep track of where they left off, so if they disconnect, they just resume where they should, or they can decide to rewind. [ editor - I decided I hated my description. Retrying. ]
This is a terrible description of some awesome software, which I will point out is open source and free, and does much, much more. But the idea is that this is for storage, not querying or logic, but it makes querying and logic in external scripts easier.
Version 0.5.2:
* Strings that can be casted to floats will get a duplicate key.num, like "temperatureF" would also get "temperatureF.num" if possible.
* Add a Unix epoch timestamp field. This makes it possible to replay out of Kafka into InfluxDB and get nearly the same information as the Influx plugin.