After starting up a big data timeseries project at work for the past several years, and spending a big chunk of that looking into and testing various DBs, whenever I am asked, "What database I should use for X?" I tell them Postgres, unless you have an well-researched and proven reason not to. I picked Druid for our project, and am very happy with the result. However, for some of our smaller customers, Postgres would have been a much better choice. Only because most of our big, important customers have data needs starting at Petabytes am I happy we went with Druid. Well, that and a bunch of other reasons ;)

Much of the "reasoning" I see for people choosing Elasticsearch or InfluxDB or MongoDB seems to come down to, "It showed up a lot on Hackernews".