Graphite / StatsD push ingestion method for Druid?

There appears to be a dearth of support for ingesting time-series data into Druid via the StatsD protocol or Graphite protocol. I can only assume it’s not supported, but is it necessarily not so? As in, would it be feasible to build a relay ingestion proxy process using the Tranquility client?

Asking as I’m looking for a way to store Kamon metrics into a time series database, and I’m evaluating several (KairoDB, InfluxDB, and Druid).

Thanks!

Tim

Hey Tim,

I don’t think a statsd or graphite bridge exists but that would totally be possible. There’s a lot of folks out there using Druid for time series so it’s definitely something that makes sense. (for example, at the last Druid meetup, PayPal shared some details about their use of it in their monitoring platform)

If you’re already using Kafka to ship metrics around then Tranquillity Kafka (https://github.com/druid-io/tranquility/blob/master/docs/kafka.md) is probably the easiest way to go. Otherwise, try Tranquility Core if you like writing Java/Scala programs or Tranquility Server if you don’t.