Druid to lookup on Dimension


I’m using druid OLAP use case. I’m trying get info on the possibility of using druid to lookup on a dimension.

Metadata: Dataset is segmented by hour and each hour generates of 500MB. I don’t have any limitation of size of cluster. And let’s assume that network adds 0 ms to response time.

With tuning is it possible to achieve response within 60ms with about 20,000 rps? I’m planning to use “select” query. I know exact time(range won’t be greater than 2 hours) of the dimension before querying druid. Has anyone done any thing similar? Will druid work for this use case?


Hi Prashanth,

The case you describe is not really what Druid is designed for. You would probably be happier using a key/value store designed for high read rates of single keys.