Dataloop processes and stores millions of individual time series metric paths at per second resolution to help our customers build and run successful SaaS products. Our initial data pipeline was custom built with NodeJS and Riak but this wasn't scaling in a reliable way. Six months ago we migrated our backend worker processes to Erlang to improve the reliability of our metrics and alerts processing. More recently we have switched to a new custom built Erlang time series database called Dalmatiner DB. This talk is about our experiences with Erlang so far.
Steven Acreman is a career-long DevOps person who started off supporting a small team of Developers in Watford and ended up building and running internal platforms for about 1200 software developers in a $1bn software company. More recently he helped to launch the Alfresco Cloud SaaS product and then co-founded Dataloop to solve a lot of the monitoring problems experienced over the last 10+ years.