Blog

My name is Raj Dutt, and I am the co-founder and CEO of raintank, Inc. As many of you know, raintank is the company behind Grafana, which our co-founder Torkel created as a front-end for Graphite almost 3 years ago. Today, Grafana supports dozens of different time series databases. Overall, we think Graphite is pretty awesome. Originally written by Chris Davis at Orbitz in 2006, it has helped many thousands of companies achieve a quantum leap in how they handled their metrics.

My co-founder Anthony wrote about our promising initial experiences with InfluxDB a few months ago. We’re bullish on InfluxDB, but it’ll probably take a little while for Paul and his burgeoning team to reach a production-worthy release, especially when you consider that 0.9 is a ground-up rewrite. Another promising database that we’ve been taking a hard look at in the last few weeks is KairosDB, which is a fork of OpenTSDB, but using Cassandra instead of HBase.

As we begin to develop the raintank platform, we’ve started to play with numerous solutions for time series databases including MongoDB, OpenTSDB, Graphite, and even MySQL. The more generic database solutions, MongoDB and MySQL, were quickly ruled out as viable options. Though they are both great tools, they are not suited to the narrow use case of large scale time-series data, which is of particular interest to us. OpenTSDB showed a lot of promise, but we found the complexity of keeping a Hadoop cluster running not worth the rewards.